US EPA Office of Research and Development
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-00/096
October 2000
Volatilization Rates from
Water to Indoor Air
Phase II
-------
-------
EPA 600/R-00/096
October 2000
VOLATILIZATION RATES FROM WATER TO INDOOR AIR
PHASE II
National Center for Environmental Assessment-Washington Office
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
Printed on Recycled Paper
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
ABSTRACT
Contaminated water can lead to volatilization of chemicals to residential indoor air.
Previous research has focused on only one source (shower stalls) and has been limited to
chemicals in which gas-phase resistance to mass transfer is of marginal significance. As a result,
attempts to extrapolate chemical emissions from high-volatility chemicals to lower volatility
chemicals, or to sources other than showers, have been difficult or impossible.
In this study two-phase dynamic mass balance models were developed for estimating
chemical emissions from washing machines, dishwashers, and bathtubs. An existing model was
adopted for showers only. Source- and chemical-specific mass transfer coefficients, as well as
ak exchange (ventilation) rates were estimated based on a series of experiments. These
experiments were conducted using 5 tracer chemicals (acetone, ethyl acetate, toluene,
ethylbenzene, and cyclohexane) and 4 sources (showers, bathtubs, washing machines, and
dishwashers). Each set of experiments led to the determination of chemical stripping efficiencies
and mass transfer coefficients (overall, liquid-phase, gas-phase), and to an assessment of the
importance of gas-phase resistance to mass transfer. Stripping efficiencies ranged from 6.3% to
80% for showers, 2.6% to 69% for bathtubs, 18% to 100% for dishwashers, and 3.8% to 100%
for washing machines. Acetone and cyclohexane always defined the lower and upper bounds,
respectively, of these ranges.
11
-------
CONTENTS
List of Tables
List of Figures xi
Nomenclature and Abbreviations xiii
Preface xvj
Authors, Contributors, and Reviewers xvii
Executive Summary xix
1. Introduction 7. ... 1-1
1.1. Problem Statement ... 1_1
1.2. Research Objectives i_2
1.3. Scope of Research : 1-2
1.4. Organization of Research Report 1-2
2. Model Development 2-1
2.1. Mass Transfer Theory , 2-1
2.1.1. Chemical Stripping Efficiency 2-1
2.1.2. Mass Transfer Coefficients 2-2
2.2. Ideal Reactor Models 2-9
2.2.1. Plug Flow Reactor Model 2-9
2.2.2. Continuous-Flow Stirred-Tank Reactor Model ( 2-10
2.2.3. BatchReactorModel 2-11
2.3. Source-Specific Mass Balance Models 2-12
2.3.1. Dishwasher Models .-....• 2-12
2.3.2. Washing Machine Models 2-16
2.3.2.1. Washing Machine Fill Cycle 2-16
2.3.2.2. Washing Machine Wash/Rinse Cycles ...... . 2-18
2.3.3. Shower Models '....,.. 2-18
2.3.4. Bathtub Models 2-21
2.3.4.1. Bathtub Flow-Through Model 2-21
2.3.4.2. Bathtub Fill Model ., 2-21
2.3.4.3. Bathtub Surface Volatilization Model , 2-21
2.4. Chemical Emission Models 2-22
111
-------
CONTENTS (continued)
3. General Methodology 3-1
3.1. Source Chamber 3-1
3.2. Chemical Tracers 3-2
3.2.1. Physicochemical Properties 3-2
3.2.2. Chemical Tracer Addition 3-3
3.3. Chemical Sampling .. : 3-4
3.3.1. Liquid-Phase Sampling ...." 3-4
3.3.2. Gas-Phase Sampling 3-5
3.4. Sample Analyses 3-6
3.4.1. Liquid Sample Analysis 3-6
3.4.2. Liquid Standards 3-7
3.4.3. Gas Sample Analysis 3-9
3.4.4. Gas Standards ......:.... 3-9
3.5. Quality Assurance Measures 3-11
3.5.1. Duplicate Samples 3-11
3.5.2. Replicate Experiments 3-12
3.5.3. Experimental Blanks .-..' 3-13
3.5.4. Method Detection Limit 3-13
3.6. Data Analysis 3-14
3.6.1. Chemical Stripping Efficiencies ..'...• 3-15
3.6.2. Overall Mass Transfer Coefficients (KLA) 3-15
3.6.3. Ratio of Gas-to-Liquid Phase Mass Transfer Coefficients 3-17
3.6.4. Liquid- and Gas-Phase Mass Transfer Coefficients 3-18
3.7. Factorial Analysis 3-18
3.8. Mass Closure Assessment 3-18
4. Shower Stall Experiments 4-1
4.1. Experimental System 4-1
4.2. Experimental Design 4-3
4.3. Source-Specific Methodology 4-4
4.3.1. Sample Schedule : I 4-4
4.3.2. Ventilation Rate 4-5
4.3.3. Parameter Estimation .. 4-5
4.4 Shower Results 4-5
4.4.1. Chemical Stripping Efficiencies 4-5
IV
-------
CONTENTS (continued)
4.4.2. KLA Values 4-17
4.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients 4-26
4.4.4. Mass Closure 4-30
5. Dishwasher Experiments . 5-1
5.1. Experimental System .5-1
5.2. Experimental Design 5-2
5.3. Source-Specific Methodology 5-3
5.3.1. Sample Schedule ,,..." 5-4
5.3.2. Ventilation Rate : 5-4
5.3.3. Parameter Estimation 5-4
5.4. Dishwasher Results 5-7
5.4.1. Ventilation Rates . 5-7
5.4.2. Chemical Stripping Efficiencies 5-8
5.4.3. KLA Values 5-12
5.4.4. Liquid- and Gas-Phase Mass Transfer Coefficients 5-19
5.4.5. Mass Closure 5-20
6. Washing Machine Experiments ............... . . 6-1
6.1. Fill Cycle Experiments 6-1
6.1.1. Experimental System 6-1
6.1.2. Experimental Design 6-3
6.1.3. Source-Specific Methodology 6-3
6.1.3.1. Sample Schedule .. ........ 6-4
6.1.3.2. Ventilation Rates .. 6-5
6.1.3.3. Parameter Estimation . . '... 6-6
6.1.4. Fill Cycle Results 6-7
6.1.4.1. Ventilation Rates . ,. 6-8
6.1.4,2. Chemical Stripping Efficiencies 6-9
6.1.4.3. KLAValues 6-11
6.1.4.4. Liquid- and Gas-Phase Mass Transfer Coefficients 6-12
6.1.4.5. Mass Closure . . . . . ...... ..-..' 6-13
6.2. Wash/Rinse Cycle Experiments ................ • 6-14
6.2.1. Experimental System .:........ 6-14
6.2.2. Experimental Design 6-15
-------
CONTENTS (continued)
6.2.3. Source-Specific Methodology 6-15
6.2.3.1. Sample Schedule 6-16
6.2.3.2. Ventilation Rates .. ... 6-16
6.2.3.3. Parameter Estimation 6-16
6.2.4. Wash/Rinse Cycle Results 6-17
6.2.4.1. Ventilation Rates 6-18
6.2.4.2. Chemical Stripping Efficiencies 6-20
6.2.4.3. KLA Values 6-31
6.2.4.4. Liquid- and Gas-Phase Mass Transfer Coefficients 6-42
6.2.4.5. Mass Closure 6-45
7. Bathtub Experiments 7-1
7.1. Bathtub Flow-Through Experiments 7-1
7.1.1. Experimental System . 7-1
7.1.2. Experimental Design .. 7-1
7.1.3. Source-Specific Methodology 7-1
7.1.3.1. Sample Schedule ........ 7-1
7.1.3.2. Ventilation Rate 7-2
7.1.3.3. Parameter Estimation 7-2
7.1.4. Bathtub Flow-Through Results 7-3
7.1.4.1. Chemical Stripping Efficiencies 7-3
7.1.4.2. KLA Values 7-5
7.1.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients 7-7
7.1.4.4. Mass Closure 7-8
7.2. Bathtub Fill Experiments 7-9
7.2.1. Experimental System •. 7-9
7.2.2. Experimental Design 7-9
7.2.3. Source-Specific Methodology 7-9
7.2.3.1. Sample Schedule . , .- 7-9
7.2.3.2. Ventilation Rates 7-11
7.2.3.3. Parameter Estimation 7-11
7.2.4. Bathtub Fill Results 7-11
7.2.4.1. Chemical Stripping Efficiencies ..7-11
7.2.4.2. Values 7-13
7.2.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients 7-13
VI
-------
CONTENTS (continued)
7.2.4.4. Mass Closure 7-13
7.3. Surface Volatilization Experiments 7-14
7.3.1. Experimental System 7-14
7.3.2. Experimental Design 7-15
7.3.3. Source-Specific Methodology 7-15
7.3.3.1. Sample Schedule 7-15
7.3.3.2. Ventilation Rates 7-15
7.3.3.3. Parameter Estimation 7-15
7.3.4. Bathtub Surface Volatilization Results .7-16
7.3.4.1. Chemical Stripping Efficiencies 7-17
7.3.4.2. KLA Values .". 7-18
7.3.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients 7-19
7.3.4.4. Mass Closure . 7-19
8. Model Applications 8-1
8.1. Shower Model Application 8-1
8.2. Dishwasher Model Application 8-7
8.3. Washing Machine Model Application •. • • • i • • • • 8-13
8.4. Bathtub Model Application . . . . 8-19
9. Summary and Conclusions 9-1
9.1. Summary 9-1
9.2. Conclusions: General 9-3
9.3. Conclusions: Showers 9-4
9.4. Conclusions: Dishwashers 9-5
9.5. Conclusions: Washing Machines . . 9-5
9.6. Conclusions: Bathtubs '...... 9-6
9.7. Recommendations for Future Research 9-7
10. References 10-1
Appendix: Chemical Volatilization Database A-l
vn
-------
LIST OF TABLES
3-1 Summary of physicochemical properties for selected chemical tracers 3-2
3-2 Duplicate sample results 3-11
3-3 Replicate sample results .'....• 3-12
3-4 Replicate sample results excluding replicate experiments associated with filling 3-13
3-5 Method detection limits (MDLs) for liquid and gas samples 3-14
4-1 Shower experiment operating conditions 4-6
4-2 Acetone stripping efficiencies for experimental shower 4-7
4-3 Ethyl acetate stripping efficiencies for experimental shower 4-7
4-4 Toluene stripping efficiencies for experimental shower 4-8
4-5 Ethylbenzene stripping efficiencies for experimental shower 4-8
4-6 Cyclohexane stripping efficiencies for experimental shower 4-9
4-7 Acetone KLA values for experimental shower . 4-18
4-8 Ethyl acetate KLA values for experimental shower 4-18
4-9 Toluene KLA values for experimental shower 4-19
4-10,EthylbenzeneKLA values for experimental shower 4-19
4-11 Cyclohexane KLA values for experimental shower 4-20
4-12 Liquid- and gas-phase mass transfer coefficients for shower experiments 4-27
5-1 Dishwasher experimental operating conditions 5-7
5-2 Dishwasher ventilation rate experimental results 5-9
5-3 Chemical stripping efficiencies (r\) for experimental dishwasher 5-10
5-4 Acetone KLA values for dishwasher experiments 5-13
5-5 Toluene KLA values for dishwasher experiments 5-13
5-6 Ethylbenzene KLA values for dishwasher experiments 5-14
5-7 Cyclohexane KLA values for dishwasher experiments 5-14
6-1 Washing machine fill cycle experimental conditions 6-8
6-2 Washing machine fill cycle ventilation rates 6-9
6-3 Chemical stripping efficiencies (r|) for washing machine fill cycles 6-10
6-4 Values of KLA for washing machine fill cycles 6-12
6-5 Liquid- and gas-phase mass transfer coefficients for washing machine fill cycle
experiments 6-13
6-6 Washing machine wash/rinse cycle experimental operating conditions 6-18
6-7 Ventilation rate experimental results .".. 6-19
6-8 Acetone stripping efficiencies for washing machine wash/rinse cycle—Factorial #1 .. 6-20
6-9 Acetone stripping efficiencies for washing machine wash/rinse cycle—Factorial #2 .. 6-21
vm
-------
LIST OF TABLES (continued)
6-10 Ethyl acetate stripping efficiencies for washing machine wash/rinse cycle—
Factorial #2. . ... 5.21
6-11 Toluene stripping efficiencies for washing machine wash/rinse cycle—Factorial #1 .. 6-22
6-12 Toluene stripping efficiencies for washing machine wash/rinse cycle—Factorial #2 .. 6-22
6-13 Ethylbenzene stripping efficiencies for washing machine wash/rinse cycle—
Factorial #1 6-23
6-14 Ethylbenzene stripping efficiencies for washing machine wash/rinse cycle— •
Factorial #2 6-23
6-15 Cyclohexane stripping efficiencies for washing machine wash/rinse cycle—
Factorial #1 6-24
6-16 Cyclohexane stripping efficiencies for washing machine wash/rinse cycle—
Factorial #2 6-24
6-17 Acetone KLA values for washing machine wash/rinse cycle—Factorial #1 6-32
6-18 Acetone KLA values for washing machine wash/rinse cycle—Factorial #2 ......... 6-32
6-19 Ethyl acetate KLA values for washing machine wash/rinse cycle—Factorial #2 6-33
6-20 Toluene KLA values for washing machine wash/rinse cycle—Factorial #1 6-33
6-21 Toluene KLA values for washing machine wash/rinse cycle—Factorial #2 6-34
6-22 Ethylbenzene KLA values for washing machine wash/rinse cycle—Factorial #1 6-34
6-23 Ethylbenzene KLA values for washing machine wash/rinse cycle—Factorial #2 6-35
6-24 Cyclohexane KLA values for washing machine wash/rinse cycle—Factorial #1 6-35
6-25 Cyclohexane KLA values for washing machine wash/rinse cycle—Factorial #2 6-36
6-26 Liquid- and gas-phase mass transfer coefficients for washing machine
wash/rinse cycle experiments—Factorial #1 6-46
6-27 Liquid- and gas-phase mass transfer coefficients for washing machine
wash/rinse cycle experiments—Factorial #2 6-47
7-1 Bathtub flow-through operating conditions 7-3
7-2 Chemical stripping efficiencies (TI) for experimental bathtub flow-through
experiments 7-4
7-3 Values of KLA for bathtub flow-through experiments .7-5
7-4 Liquid- and gas-phase mass transfer coefficients for bathtub flow-through
experiments 7.7
7-5 Bathtub (fill) operating conditions ..7-12
7-6 Chemical stripping efficiencies (T|) for bathtub (fill) experiments 7-12
7-7 Values of KLA for bathtub (fill) experiments .7-13
7-8 Liquid- and gas-phase mass transfer coefficients for bathtub (fill) experiments ...... 7-14
IX
-------
LIST OF TABLES (continued)
7-9 Bathtub surface volatilization operating conditions . ...... ..... . ..... .... ...... 7-17
7-10 Chemical stripping efficiencies for bathtub surface volatilization experiments ....... 7-18
7-11 Values of KLA for bathtub surface volatilization experiments ..... ............... 7-18
7-12 Liquid- and gas-phase mass transfer coefficients for bathtub surface volatilization
experiments ....... .......... ........................ . . . ..... . . . ...... . 7-20
8-1 Comparison of the three chemicals used in model applications .......... ;......... 8-2
9-1 Summary of experimental stripping efficiencies and kg/kj ......... . .............. 9-2
-------
LIST OF FIGURES
2-1 Plug flow reactor 2-10
2-2 Continuous-flow stirred-tank reactor 2-11
2-3 Batch reactor . 2-11
2-4 Dishwasher model 2-13
2-5 Washing machine fill cycle model 2-17
2-6 Washing machine wash/rinse cycle model 2-18
2-7 Shower model 2-19
2-8 Bathtub flow-through model 2-22
2-9 Bathtub fill model , 2-22
2-10 Bathtub surface volatilization model 2-23
3-1 Gas sampling experimental set up 3.5
3-2 Liquid-phase sample chromatogram 3.7
3-3 Liquid-phase calibration curve for ethylbenzene 3-8
3-4 Gas sample chromatogram 3_10
3-5 Gas-phase calibration curve for acetone . .-...'... 3-10
3-6 Matrix format used to determine kg/k, 3-19
4-1 Shower experimental system 4.2
4-2 Shower factorial experimental design . 4.3
Relationship between Henry's law constant and average stripping efficiency 4-16
4-4 Acetone experimental data for Experiment 7 4-20
4-5 Ethyl acetate experimental data for Experiment 7 4-22
4-6 Toluene experimental data for Experiment 7 4-23
4-7 Ethylbenzene experimental data for Experiment 7 4-24
4-8 Cyclohexane experimental data for Experiment 7 4-25
4-9 Resistances to mass transfer for each chemical in Experiment 7 4-29
5-1 Dishwasher experimental system ...... 5.2
5-2 Factorial experimental design for dishwasher experiments 5-3
5-3 Isobutylene decay due to ventilation for Experiment 18 5-9
5-4 Comparison of measured Cg/Q predicted Henry's law constant for acetone 5-16
5-5 Toluene concentrations for Experiment 8 5-17
5-6 Amplification of Figure 30 to illustrate approach to equilibrium
conditions fortoluene 5_18
5-7 Ethylbenzene concentrations for Experiment 8 5-19
5-8 Cyclohexane concentrations for Experiment 8 5-20
XI
-------
LIST OF FIGURES (continued)
6-1 Washing machine fill cycle experimental system 6-2
6-2 Isobutylene decay due to ventilation for Experiment 13 6-9
6-3 Wash/rinse cycle experimental system 6-14
6-4 Wash/rinse cycle factorial experimental design 6-15
6-5 Isobutylene decay due to ventilation for Experiment 8 6-20
6-6 Acetone concentrations for Experiment 6 6-37
6-7 Amplification of Figure 6-6 for acetone gas-phase data 6-38
6-8 Toluene concentrations for Experiment 6 6-40
6-9 Magnification of Figure 6-8 to illustrate toluene's gas-phase
concentration over tune • • 6-40
6-10 Ethylbenzene concentrations for Experiment 6 6-42
6-11 Cyclohexane concentrations for Experiment 6 6-43
6-12 Liquid- and gas-phase resistances to mass transfer for Experiment 6 6-44
7-1 Bathtub flow-through experimental system 7-2
•7-2 Toluene experimental data for Experiment 4 replicate 7-6
7-3 Resistances to mass transfer for each chemical in Experiment 2 7-8
7-4 Bathtub fill experimental system 7-10
7-5 Toluene experimental data for Experiment 4 replicate 7-19
8-1 Mass emission rates for three chemicals for example shower event 8-2
8-2 Mass emission rates for three chemicals for example dishwasher event 8-11
8-3 Amplification of Figure 52 to show methyl ethyl ketone mass emission rate 8-12
8-4 Mass emission rates for three chemicals for example washing machine event 8-18
8-5 Mass emission rates for three chemicals for example bathtub event 8-20
xn
-------
NOMENCLATURE AND ABBREVIATIONS1
A
ACH
AA
C
Cin
Q
Q,end
Q,in
m
CFSTR
cosh
coth
DBCM
DBCP
FID
GC
interfacial surface area between water and adjacent air (L2)
air changes per hour
differential 'area (L2)
chemical concentration (M/L3)
experimentally measured liquid- and gas-phase concentrations (M/L3)
chemical concentration in air adjacent to water (M/L3)
inlet concentration of contaminant in air (M/L3)
initial chemical concentration in gas volume (M/L3)
chemical concentration in air at any time t (M/L3)
inlet chemical concentration (M/L3)
chemical concentration in water (M/L3)
final chemical concentration in water (M/L3)
inlet chemical concentration in water (M/L3)
outlet concentration of contaminant in water (M/L3)
initial chemical concentration in water (M/L3)
mathematically predicted liquid- and gas-phase concentrations (M/L3)
continuous flow stirred tank reactor
hyperbolic cosine
hyperbolic cotangent
molecular diffusion coefficient for a chemical in air (L2/T)
molecular diffusion coefficient for chemical i in air (L2/T)
molecular diffusion coefficient for chemical j in air (L2/T)
molecular diffusion coefficient for a contaminant in water (L2/T)
molecular diffusion coefficient for chemical i in water (L2/T)
molecular diffusion coefficient for chemical j in water (L2/T)
dibromochloromethane
1,2-dibromo-3 -chloropropane
chemical mass emission rate (M/T)
flame ionization detector
gas chromatography
1 Note: Terms in parentheses denote units; M corresponds to mass; L corresponds to
length; T corresponds to time; (°) corresponds to temperature; dimensionless values are denoted
as(-).
xin
-------
H,T
ID
mc
MDL
MEK
OD
P
pFR
Pv
Q
Q8
Qin
Q,
Qout
sinh
NOMENCLATURE AND ABBREVIATIONS (continued)
Henry's law constant (L3gas/L3iiq)
Henry's law constant for chemical i (L3gas/L3Iiq)
Henry's law constant for chemical j (L3gas/L3liq)
Henry's law constant at experimental temperature (L3gas/L3liq)
inside diameter (L)
gas-phase mass transfer coefficient (L/T)
gas-phase mass transfer coefficient for chemical i (L/T)
gas-phase mass transfer coefficient for chemical j (L/T)
overall mass transfer coefficient for contaminant of interest (L/T)
liquid-phase mass transfer coefficient (L/T)
overall mass transfer coefficient for chemical i (L/T)
liquid-phase mass transfer coefficient for chemical i (L/T)
overall mass transfer coefficient for chemical j (L/T)
liquid-phase mass transfer coefficient for chemical j (L/T)
degree of mass closure (-)
method detection limit
methyl ethyl ketone
power constant for ratio of liquid-phase diffusion coefficients (-)
power constant for ratio of gas-phase diffusion coefficients (-)
outside diameter (L)
perimeter (L)
plug flow reactor
vapor pressure (L Hg)
volumetric flowrate (L3/T)
gas flowrate (L3/T)
inlet volumetric flowrate (L3/T)
liquid flowrate (L3/T)
outlet volumetric flowrate (L3/T)
area reaction rate (M/L2«T)
rate of surface renewal for the gas side of the interface (1/T)
rate of surface renewal for the liquid side of the interface (1/T)
volume reaction rate (M/L3«T)
hyperbolic sine
standard deviation of replicate analyses
xiv
-------
t
T
Tb
TCE
TKE
V
AV
V
V,
z
80
NOMENCLATURE AND ABBREVIATIONS (continued)
time (T)
temperature (°C)
boiling point (°C)
trichloroethene
total kinetic energy
volume (L3)
differential volume (L3)
local volume of air (L3)
local volume of water (L3)
direction of flow
thickness of a hypothetical gas film adjacent to the interface and
through which contaminant transport is solely by molecular diffusion (L)
thickness of a hypothetical liquid film adjacent to the interface and through
which contaminant transport is solely by molecular
diffusion (L)
chemical stripping efficiency (-)
density (M/L3)
gas-phase mass transfer relational coefficient (-)
liquid-phase mass transfer relational coefficient (-)
overall mass transfer relational coefficient (-)
xv
-------
PREFACE
This report was prepared under the direction of the National Center for Environmental
Assessment (NCEA) of EPA's Office of Research and Development (ORD). The purpose of
this report is to provide a methodology for estimating chemical emissions from washing
machines, dishwashers, showers, and bathtubs. The methodology presented in this report was
derived from volatilization experiments conducted by The University of Texas at Austin under a
Cooperative Agreement with NCEA. The results of these experiments are included in the report.
The report was submitted in fulfillment of Cooperative Agreement No. CR 824228-01
and covers the period from June 1, 1995 to August 31, 1997, and work was completed as of
August 31,1997.
XVI
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
This report was prepared under Cooperative Agreement No. CR 824228-01 between The
University of Texas at Austin and the National Center for Environmental Assessment (NCEA),
Office of Research and Development. Jacqueline Moya was responsible for the overall
coordination, direction, and technical assistance.
AUTHORS
Cynthia Howard-Reed
formerly Graduate Research Assistant with The University of Texas at Austin
presently with the Indoor Air Quality Group at the National Institute of Standards and
Technology (NIST)
Richard L. Corsi
Principal Investigator
Environmental and Water Resources Engineering Program
The University of Texas at Austin
REVIEWERS
The following individuals have reviewed this report and provided valuable comments:
Environmental Protection Agency Reviewers
Jacqueline Moya
Nancy Chiu
Lance Wallace
External Reviewers
John Little
Associate Professor
Civil and Environmental Engineering
Virginia Tech
Blacksburg, VA
Nicholas J. Giardino
Toxicologist
Brooks AFB, Texas
xvn
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
ACKNOWLEDGMENTS
The authors wish to thank EPA Project Officer Jacqueline Moya for her general guidance
and enthusiasm regarding this project. The authors also wish to acknowledge Albert Chung,
Jennifer Pettibon, Javier Ramirez, Tony Smith, and Ross Strader, undergraduate students at The
University of Texas at Austin, for their assistance during experiments.
xvin
-------
EXECUTIVE SUMMARY
Contaminated water can lead to volatilization of chemicals to residential indoor air.
Previous research has focused on only one source (shower stalls) and has been limited to
chemicals in which gas-phase resistance to mass transfer is of marginal significance. As a result,
attempts to extrapolate chemical emissions from high-volatility chemicals to lower volatility
chemicals, or to sources other than showers, have been difficult or impossible.
In this study two-phase dynamic mass balance models were developed for estimating
chemical emissions from washing machines, dishwashers, and bathtubs. An existing model was
adopted for showers only. The mass transfer theory and derivations of these models are further
described in chapter 2 of this report. Source- and chemical-specific mass transfer coefficients, as
well as air exchange (ventilation) rates were estimated based on a series of experiments. These
experiments were conducted using 5 tracer chemicals (acetone, ethyl acetate, toluene,
ethylbenzene, and cyclohexane) and 4 sources (showers, bathtubs, washing machines, and
dishwashers). Each set of experiments led to the determination of chemical stripping efficiencies
and mass transfer coefficients (overall, liquid-phase, gas-phase), and to an assessment of the
importance of gas-phase resistance to mass transfer. , -
A set of protocols for estimating emission rates for chemicals other than those used in this
study was defined for each of the four sources. Example applications are provided and illustrate
the dynamic behavior of emissions and importance of chemical properties on such emissions.
The experimental mass transfer coefficients, air exchange rates and protocols described in this
report can be used as direct input values or to estimate reasonable input values for the reported
emission models.
Stripping efficiencies ranged from 6.3% to 80% for showers, 2.6% to 69% for bathtubs,
18% to 100% for dishwashers, and 3.8% to 100% for washing machines. Acetone and
cyclohexane always defined the lower and upper bounds, respectively, of these ranges.
The findings of this study lead to several conclusions. A detailed discussion of
conclusions is presented in chapter 9. Some of the most significant conclusions are summarized
below.
System operating conditions can have a significant effect on chemical emissions. In
particular, chemical stripping efficiencies for washing machines were observed to be
highly sensitive to system operating conditions.
Water temperature was an important variable that affected stripping efficiencies and mass
transfer coefficients for all sources.
xix
-------
Chemical stripping efficiencies increase as Henry's law constant increases for lower-
volatility chemicals. However, with the exception of the fill-cycle of bathtubs, chemical
stripping efficiencies are relatively insensitive to Henry's law constant for chemicals with
constants greater than that of toluene.
Failure to account for gas-phase resistance to mass transfer can lead to significant
overestimates of chemical volatilization to indoor air. This is particularly true for lower-
volatility chemicals or those sources with low values of gas- to liquid-phase mass transfer
coefficients (kg/k,), e.g., washing machines.
Results for shower experiments were reasonably consistent with those reported by other
researchers with stripping efficiencies ranging from 60% to 80% for chemicals with
Henry's law constant equal or greater than that of toluene.
Gas-phase concentrations were homogeneous throughout the shower stall demonstrating
that the frequent assumption of a well-mixed system is reasonably accurate.
Dishwashers were determined to be very effective at removing chemicals from water to
air, with low but continuous emissions during operation and significant storage within the
dishwasher headspace. The most significant release of chemicals to indoor air would
occur if the dishwasher door is opened immediately after use.
Washing machines during the rinse cycle with hot water and low clothes loading resulted
in stripping efficiencies that approached 100% for chemicals with Henry's law constant
greater than toluene.
Bathtubs may be more significant than showers with respect to human exposure to
chemicals dissolved in water because of longer exposure times.
xx
-------
1. INTRODUCTION
1.1. PROBLEM STATEMENT
Current Federal drinking water (chemical) standards are primarily based on associated
ingestion exposure. However, other exposure routes—inhalation and dermal contact—may be as
or more important in terms of human health risk (McKone, 1987). Contaminated tap water can
enter a home through several sources, including showers, bathtubs, washbasins, dishwashers, and
washing machines. For each of these sources, chemicals have the potential to volatilize to indoor
air and thus provide an inhalation risk to humans. Previous studies on volatilization of chemicals
from drinking water to indoor air have been narrow in focus and have failed to close several
"knowledge gaps." These knowledge gaps currently hinder accurate inhalation exposure
assessments.
To date, research on volatilization of chemicals from drinking water has focused on one
household source, showers. Despite the relatively large number of shower experiments, previous
studies have focused on stripping efficiencies for a narrow range of chemicals, primarily radon
(Rn-222), trichloroethene (TCE), and chloroform (CHC13). Thus, current methods for estimating
inhalation exposure related to contaminated tap water are based on simplifying assumptions
and/or extrapolation techniques that fail to capture the mechanistic behavior of the volatilization
process. For example, these extrapolation techniques are limited by the lack of specific mass
transfer coefficients for chemicals that vary significantly in their volatility. Thus, there is an
important need to expand current knowledge about to chemical volatilization from tap water.
This research project was completed in two phases. The first phase was dedicated to an
extensive literature search to establish the current knowledge base regarding mass transfer of
volatile chemicals from household water sources. This literature search has been documented
elsewhere (Corsi et al, 1996). Results of Phase I led to the conclusion that there was a
significant need for experiments to estimate chemical volatilization for all household
consumptive water uses. Four sources were chosen for further study (Phase II). The four
sources were showers, dishwashers, washing machines, and bathtubs, all of which were chosen
based on predicted significance of chemical emissions.
A series of experiments was completed to determine chemical volatilization rates for each of
the four sources. Results from this work have been used to develop and evaluate improvements
1-1
-------
to existing extrapolation models for relating rates of volatilization and mass transfer coefficients
between chemicals.
1.2. RESEARCH OBJECTIVES
The research objectives of Phase II were as follows: .
1. Use a series of well-designed laboratory experiments to expand the general knowledge base
associated with chemical volatilization to indoor air.
2. Develop and evaluate improvements to existing models based on empirical and mechanistic
accounting for source operating conditions and contaminant physicochemical properties.
3. Compile and organize these experimental data in a database that may be easily used by
regulators, consultants, academics, and others.
1.3. SCOPE OF RESEARCH
A two-phase mass balance model was developed for each of the four sources described in
Section 1.1. Laboratory experiments were designed such that the mass balance model was
solved to back-calculate mass transfer coefficients for each chemical and source. These
experiments were completed using a water supply spiked with a cocktail of chemicals
representing a wide range of Henry's law constants. For each source, chemical stripping
efficiencies and mass transfer coefficients were determined for several applicable operating
conditions (water temperature, liquid flowrate, presence of detergent, etc.). A total of 113 mass
transfer and air exchange rate experiments were completed.
1.4. ORGANIZATION OF RESEARCH REPORT
Mass balance models used for each experimental source are presented in Section 2. General
experimental methodologies and analytical techniques are described in Section 3. Sections 4
through 7 include presentations of each experimental system, experimental design, source-
specific methodologies, and experimental results for the respective sources. A model application
for each source is presented in Section 8. Conclusions and recommendations based on this
research are presented in Section 9. All applicable references are included in Section 10.
Finally, the data from the experimental database are provided in the Appendix.
1-2
-------
2. MODEL DEVELOPMENT
The first step in studying each experimental source was to develop a two-phase mass
balance model describing the rate at which a chemical enters, leaves, or accumulates within the
system. A key component of each source model was the rate at which a chemical leaves the
liquid phase and enters the gas phase. This volatilization rate may be described in terms of a
stripping efficiency or mass transfer coefficient. To determine the volatilization rate for a source,
a separate mass balance was completed on each phase. By simultaneously solving the respective
differential equations, the source-specific chemical volatilization rate was predicted.
; This section is divided into four parts. Section 2.1 involves the theory of mass transfer and
important mass transfer parameters. Section 2.2 presents the different mass balance models used
for ideal reactors. Section 2.3 describes the mass balance models developed for each
experimental system. Finally, Section 2.4 describes the chemical emission models associated
with each source.
2.1. MASS TRANSFER THEORY
Although the operation of each household tap water source is very different, a similar group
of mass transfer parameters may be applied to characterize the volatilization of chemicals. These
parameters include chemical stripping efficiency, mass transfer coefficients (overall, liquid-
phase, and gas-phase), and the ratio of overall mass transfer coefficients for any two
contaminants. Each of these values is discussed in this section.
2.1.1. Chemical Stripping Efficiency
The stripping efficiency of a specific chemical for a flow-through system with a constant
volumetric fiowrate of water and no reactions is defined as:
C,
where
•'l.ou
tl=l-
= stripping efficiency (fractional) .
= outlet chemical concentration in water (M/L3)
= inlet chemical concentration in water (M/L3).
(2-1)
Similarly, the stripping efficiency of a specific chemical for a batch system with a constant
volume of water and no reaction is defined as:
2-1
-------
/,o
(2-2)
where , • '•
r\ — stripping efficiency (fractional)
C1)end = final chemical concentration in water (M/L3)
Cli0 = initial chemical concentration in water (M/L3).
In general, a stripping efficiency requires measurement of the inlet liquid concentration to
the system of interest and measurement of the outlet liquid concentration at the system's drain.
Stripping efficiency values are influenced by several factors, including chemical properties (e.g.,
Henry's law constant [Hc]), temperature, nozzle type, liquid fiowrate, gas fiowrate, and presence
of a person, detergent, clothes, and dishes (depending on the type of source).
2.1.2. Mass Transfer Coefficients
Each household source is characterized by a unique combination of mass transfer
mechanisms that affect chemical volatilization rates. These mechanisms can include a falling
film, (e.g., the jet associated with a faucet), spray droplets (e.g., in showers or dishwashers),
splashing at surfaces (e.g., during the filling of a washing machine), and entrained air bubbles
(e.g., when a faucet jet impacts an underlying basin). When two or more of these mechanisms
are important, it is often difficult or impossible to determine separate mass transfer coefficients.
It is common to "lump" the effects of multiple mass transfer mechanisms into a single overall
mass transfer coefficient, KL. The resulting equation for local rate of change of mass in the liquid
phase associated only with volatilization is:
l,out
(2-3)
where
V -
H. =
chemical concentration in water (M/L3)
local volume of water (L3)
time(T)
overall mass transfer coefficient for the chemical of interest (L/T)
contaminant concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
2-2
-------
A = interfacial surface area between water and adjacent air (L2).
For a dilute aqueous solution, the Henry's law constant for a specific chemical is defined as the
ratio of chemical concentration in air to that in water at equilibrium. Values of Henry's law
constant are dependent on chemical structure and water temperature. For most volatile organic
compounds, the Henry's law constant can be approximated closely as the ratio of chemical vapor
pressure to solubility in water. As such, a chemical (e.g., acetone) may be considered relatively
volatile in its pure state (high vapor pressure) but also relatively nonvolatile when dissolved in
water (miscible in water). The term (Q - Cg/HJ is the concentration driving force between the
liquid and gas phases. As the difference between C{ and Cg/Hc decreases, the system approaches
chemical equilibrium.
In accordance with two-film theory (Lewis and Whitman, 1924), the overall mass transfer
coefficient can be expressed as:
1 1
— = —+ .
^ *i
(2-4)
where
KL
H,
= overall mass transfer coefficient for the chemical of interest (L/T)
= liquid-phase mass transfer coefficient (L/T)
= gas-phase mass transfer coefficient (L/T)
= Henry's law constant for chemical of interest (L3liq/L3gas).
The term 1/KL is referred to as an overall resistance to mass transfer. The term 1/k, is referred to
as liquid-phase resistance to mass transfer, and l/(kg«Hc) is referred to as gas-phase resistance to
mass transfer. For kg«Hc » k,, gas-phase resistance to mass transfer is small and the overall
mass transfer coefficient is approximately equal to the liquid-phase mass transfer coefficient.
This condition is generally true for highly volatile compounds such as radon.
It is often difficult to separate mass transfer coefficients and the interfacial area (A) over
which mass transfer occurs. This is particularly true for sources without well-defined or
quiescent surfaces, for example, most indoor consumptive water uses. However, by dividing
each term by I/A, Equation 2-4 can be effectively rewritten as:
v
(2-5)
2-3
-------
where
KL
A
k,
kg
H,
= overall mass transfer coefficient for the chemical of interest (L/T)
= interfacial surface area between water and adjacent air (L2)
= liquid-^phase mass transfer coefficient (L/T)
= gas-phase mass transfer coefficient (L/T) .
= Henry's law constant for chemical of interest (L3Iiq/L3gas).
In accordance with two-film theory (Lewis and Whitman, 1924), liquid- and gas-phase mass
transfer coefficients are related to chemical properties and fluid flow conditions as follows:
(2-6)
where
Si
(2-7)
liquid-phase mass transfer coefficient (L/T)
gas-phase mass transfer coefficient (L/T)
molecular diffusion coefficient for a chemical in water (L2/T)
molecular diffusion coefficient for a chemical in air (L2/T)
thickness of a hypothetical liquid film adjacent to the interface and through which
chemical transport is solely by molecular diffusion (L)
thickness of a hypothetical gas film adjacent to the interface and through which
chemical transport is solely by molecular diffusion (L). -..-
Molecular diffusion coefficients vary to some extent between volatile chemicals and are a
function of fluid temperature. The hypothetical film thicknesses are assumed to be a function of
the extent of turbulent kinetic energy and subsequent mixing on either side of the interface,
decreasing in width with an increase in turbulent kinetic energy. For dilute aqueous solutions,
the film thicknesses are assumed to be independent of chemical concentrations.
For penetration theory (Higbie, 1935) and surface renewal theory (Danckwerts, 1951),
liquid- and gas-phase mass transfer coefficients are predicted to be proportional to the product of
molecular diffusion coefficients and surface renewal rates according to:
2-4
-------
(2-8)
where
kg =
liquid-phase mass transfer coefficient (L/T)
gas-phase mass transfer coefficient (L/T)
molecular diffusion coefficient for a chemical in water (L2/T)
molecular diffusion coefficient for a chemical in air (L2/T)
rate of surface renewal for the liquid side of the interface (1/T)
rate of surface renewal for the gas side of the interface (1/T).
(2-9)
The hypothetical surface renewal rates are assumed to be independent of contaminant
concentrations for dilute aqueous solutions. They are assumed to increase as turbulent kinetic
energy in the bulk fluid increases.
Finally, Dobbins (1956) developed a theory (film-penetration theory) that incorporates the
fundamental principles of both two-film and penetration theories. Corresponding relationships
for k, and kg are:
(2-10)
(2-11)
where
k,
liquid-phase mass transfer coefficient (L/T)
gas-phase mass transfer coefficient (L/T)
molecular diffusion coefficient for a chemical in water (L2/T)
molecular diffusion coefficient for a chemical in air (L2/T)
thiclcness of a hypothetical liquid film adjacent to the interface and through which
chemical transport is solely by molecular diffusion (L)
thickness of a hypothetical gas film adjacent to the interface and through which
chemical transport is solely by molecular diffusion (L)
2-5
-------
r, = rate of surface renewal for the liquid side of the interface (1 /T)
rg = rate of surface renewal for the gas side of the interface (1/T).
Film-penetration theory reduces to two-film theory as r, and rg become small, and to penetration
theory as TJ and rg become large.
Although values of hypothetical film thicknesses or surface renewal rates are not readily
measurable, Equations 2-6 through 2-11 are fundamental for relating mass transfer coefficients
between chemicals. Because the influences of hydrodynamic characteristics (d or r) are
independent of chemical concentrations or characteristics in dilute aqueous solutions, the ratio of
liquid-phase mass transfer coefficients for two compounds can be expressed as:
(2-12)
where
A
DU
n,
liquid-phase mass transfer relational parameter (-)
liquid-phase mass transfer coefficient for chemical i (L/T)
liquid-phase mass transfer coefficient for chemical j (L/T)
interfacial surface area between water and adjacent air (L2)
liquid-phase diffusion coefficient for chemical i (L2/T)
liquid-phase diffusion coefficient for chemical j (L2/T)
liquid-phase power constant (-).
The power constant nj varies from 0.5 (penetration and surface renewal theory) to 1.0 (two-
film theory). Values of nj have been calculated for natural and engineered systems and are often
reported to be between 0.6 and 0.7 (Roberts et al., 1984; Smith et al., 1980). A value of 2/3 is
commonly applied. Given a specific value of nt, Equation 2.12 can be used to estimate k,;, given
a measured or estimated ky and the ratio of liquid-phase diffusion coefficients between i and j.
The latter are generally available for many volatile chemicals and may also be estimated with a
reasonable degree of accuracy (Tucker and Nelken, 1990). Although DH and Dy are both
functions of temperature, researchers have observed that T, does not vary significantly with
variations in water temperature, occurrence of surfactants, or degree of turbulent mixing,
although the latter may cause variations in nj (decreasing with an increase in the amount of
turbulent kinetic energy in the water) (Matter-Muller et al., 1981; Smith et al., 1980).
2-6
-------
In a manner similar to that for ku and klj5 the ratio of gas-phase mass transfer coefficients for
two compounds can be expressed as:
(2-13)
where
Yg = gas-phase mass transfer relational parameter (-)
kgi = gas-phase mass transfer coefficient for chemical i (L/T)
kgj = gas-phase mass transfer coefficient for chemical j (L/T)
Dgi = gas-phase diffusion coefficient for chemical i (L2/T)
Dg = gas-phase diffusion coefficient for chemical j (L2/T)
n2 = power constant (-).
As with the power constant nl5 n2 varies between 0.5 and 1.0. Relative to studies involving
estimates of nl5 less work has been reported to confirm appropriate values of n2. A value of n2
equal to 2/3 is often assumed (Little, 1992).
In accordance with Equation 2-4, the ratio of overall mass transfer coefficients for chemicals
i and j can be expressed as: •
1111
—+ —-+
K
LJ
(2-14)
where
KLi
A
Hci =
Hcj
overall mass transfer coefficient for chemical i (L/T)
overall mass transfer coefficient for chemical j (L/T)
interfacial surface area between water and adjacent air (L2)
liquid-phase mass transfer coefficient for chemical i (L/T)
liquid-phase mass transfer coefficient for chemical j (L/T)
gas-phase mass transfer coefficient for chemical i (L/T)
gas-phase mass transfer coefficient for chemical j (L/T)
Henry's law constant for chemical i (L3liq/L3gas)
Henry's law constant for chemical j (L3liq/L3gas).
2-7
-------
Assuming kg/kg is equal to kgi/kli? Equation 2-14 can be rearranged algebraically to yield:
1 +
(2-15)
where
KLi
= overall mass transfer coefficient relational parameter (-)
= overall mass transfer coefficient for chemical i (L/T)
= overall mass transfer coefficient for chemical j (L/T)
T, = liquid-phase mass transfer relational parameter (-)
*L\
A
Dj; = liquid-phase diffusion coefficient for chemical i (L2/T)
By = liquid-phase diffusion coefficient for chemical j (L2/T)
nt = liquid-phase power constant (-)
Yg = gas-phase mass transfer relational parameter (-) = —^
gas-phase diffusion coefficient for chemical i (L2/T)
gas-phase diffusion coefficient for chemical j (L2/T)
power constant (-)
Henry's law constant for chemical i (L3Iiq/L3gas)
n2
Hci
Henry's law constant for chemical j (L3liq/L3gas)
liquid-phase mass transfer coefficient for chemical j (L/T)
gas-phase mass transfer coefficient for chemical j (L/T).
A common mistake when relating overall mass transfer coefficients between two chemicals
is to assume that KLi/KLj = Tj. This relationship requires knowledge only of liquid molecular
diffusion coefficients for each compound in accordance with Equation 2-12, but is valid only
when gas-phase resistance to mass transfer is negligible for each compound. In fact, Equation 2-
2-8
-------
15 converges to Equation 2-12 as kg/k, and/or Hcfor both i and j become very large. As discussed
previously, an assumption that gas-phase resistance is negligible is reasonable when both
compounds are highly volatile (e.g., radon). However, for less volatile compounds, it may be
necessary to know or estimate not only liquid-phase molecular diffusion coefficients, but also
gas-phase molecular diffusion coefficients, Henry's law constants for each chemical, and the
ratio of gas-to-liquid phase mass transfer coefficients for the relational surrogate j . Diffusion
coefficients and Henry's law constants can be readily obtained or estimated for most chemicals.
However, there has not been a significant amount of published information related to kg/kj for
indoor sources, including its variability with source operating conditions (e.g., liquid flowrate).
Thus, an objective of this project was to determine values of kgA, kjA, and kg/k[ for a wide range
of operating conditions associated with each experimental system.
2.2. IDEAL REACTOR MODELS
Three ideal reactor models are often used to simulate aqueous and gaseous systems. First, a
system may behave as an ideal plug flow reactor (PFR) in which fluid parcels move in an orderly
manner without any contact of other fluid parcels in the axial direction, that is, no axial
dispersion. A second ideal reactor is a continuous-flow stirred-tank reactor (CFSTR), where
fluid parcels are completely mixed within the system such that the concentration within the
reactor is the same at all locations. The influent stream of a CFSTR is instantaneously mixed
with fluid in the reactor, and the exit stream has the same concentration as fluid within the
reactor. Finally, an ideal batch reactor may be used to describe systems where fluid is initially
introduced to a system as a well-mixed uniform solution, after which no fluid enters or leaves the
reactor. Most household water systems have behavior that falls between these ideal cases.
However, instead of developing nonideal flow models, for this project each experimental system
was analyzed as though it were ideal. Deviations from the assumed ideal case are reflected in the
experimentally determined mass transfer coefficients.
2.2.1. Plug Flow Reactor Model
A schematic of a plug flow reactor is shown in Figure 2-1. The associated mass balance
equation for a differential element of volume A V is:
(2-16)
where
dt
chemical concentration (M/L3)
2-9
-------
V
t
AA =
Q =
AV =
volume (L3)
time (T)
differential area (L2)
volumetric flowrate (L3/T)
volume reaction rate (M/L3»T)
differential volume (L3)
area reaction rate (M/L2»T)
direction of flow.
2.2.2. Continuous-Flow Stirred-Tank Reactor Model
The following equation applies to the CFSTR depicted in Figure 2-2:
where
C
V
Qin
C-
Qout
dt
/=!
jnj ~ QoM,totaiC+a ryjV + a rAJAj
(2-17)
A
chemical concentration in reactor (M/L3)
volume (L3)
inlet volumetric flowrate (L3/T)
inlet chemical concentration (M/L3)
outlet volumetric flowrate (L3/T)
volume reaction rate (M/L3«T)
area reaction rate (M/L2«T)
area (L2). /-..
+z
v
V
AV
C + dC
Az
f
-'out
Figure 2-1. Plug flow reactor.
2-10
-------
c;n
Qin
(.out
Figure 2-2. Continuous-flow stirred-tank reactor.
2.2.3. Batch Reactor Model
A batch reactor is merely a simplified CFSTR, that is, a well-mixed solution with no
volumetric flowrate terms. A mass balance on the batch reactor shown in Figure 2-3 is:
d(CV) °* r~* •
where
C
V
(2-18)
A
chemical concentration in reactor (M/L3)
volume (L3)
volume reaction rate (M/L3»T)
area reaction rate (M/L2«T)
area (L2).
V
Figure 2-3. Batch reactor.
2-11
-------
2.3. SOURCE-SPECIFIC MASS BALANCE MODELS
Development of source-specific mass balance models involved the following steps: (1)
defining the system's phase boundaries and (2) determining which ideal flow model most
accurately represented each phase. The mass balances for each source, as well as solutions to the
resulting differential equations, are given in this section.
2.3.1. Dishwasher Models
Dishwasher operation consists of pumping hot water through a rotating spray arm that
produces liquid droplets that impact surrounding surfaces. The volume of liquid used in
operation is recycled, that is, the liquid volume is constant. Typical dishwasher operation consists
of four cycles: prerinse, wash, rinse, and final rinse. Within each cycle is a fill period, the
cycling of water, and a drain period. The fill and drain periods are significantly shorter (100
seconds each) than the cycling of water, and were not modeled for this study.
Figure 2-4 represents a dishwasher, for which the liquid phase is treated as a well-mixed
batch reactor. Chemical volatilization is primarily due to the formation and spraying of droplets.
Following Equation 2-18 with the only reaction term being the transfer of mass across the
water/air interface (see Equation 2-3), a mass balance on the liquid phase leads to:
dt
•A
(2-19)
where
C,
v,
t
Hc
A
chemical concentration in water (M/L3)
volume of water (L3)
time (T)
overall mass transfer coefficient for the^chemical of interest (L/T)
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3iiq/L3gas)
interfacial surface area between water and adjacent air (L2).
2-12
-------
V,
Figure 2-4. Dishwasher model.
Assuming the liquid volume is constant during operation, the liquid-phase mass balance may be
rewritten as:
dQ=_K,A
dt V,
C.+
(2-20)
where
C,
V
t
Hc
A
chemical concentration in water (M/L3)
volume of water (L3)
time(T)
overall mass transfer coefficient for the chemical of interest (L/T)
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
interfacial surface area between water and adjacent air (L2).
The dishwasher headspace (gas phase) is assumed to approach a CFSTR, also with a single
reaction term related to mass transfer across the water/air interface. A corresponding mass
balance leads to:
dt
(2-21)
where
a
v_ =
chemical concentration in air adjacent to water (M/L3)
headspace volume (L3)
2-13
-------
t
QE
c,
Hc
A
time (T)
ventilation rate (L3/T)
gas concentration entering system from outside air (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
chemical concentration in water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
interfacial surface area between water and adjacent air (L2).
Assuming the gas volume is constant during operation and the background air is relatively clean
(Cgin = 0), Equation 2-21 may be rewritten as:
dCg. KTA
O Li
dt ~ V
Qg KrA
(2-22)
where
t
Qg
KL
C,
Hc
A
chemical concentration in air adjacent to water (M/L3)
headspace volume (L3)
tune (T)
ventilation rate (L3/T)
overall mass transfer coefficient for the chemical of interest (L/T)
chemical concentration in water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
interfacial surface area between water and adjacent air (L2).
Equations 2-20 and 2-22 must be solved simultaneously in order to determine KLA.
Z
'T-'
exp - —f|sinh] |
V^
-E\t\
Analysis using Laplace transforms leads to:
•—/I cos
2
I^-'I'I
and
(2-23)
2-14
-------
C^H-f'lHNT""*1'1*
DC
where
C
IO = initial liquid concentration (M/L3)
g o = initial gas concentration (M/L3)
= KLA
(2-24)
KL
A
V]
overall mass transfer coefficient for the chemical of interest (L/T)
interfacial surface area between water and adjacent air (L2)
volume of water (L3)
B =
Hc = Henry's law constant for chemical of interest (L3,iq/L3gas)
X =
K.A
Li
Vg = headspace volume (L3)
Y =
Qg = ventilation rate (L3/T)
D = Z + Y
E = ZY-BX
F = ZCg.o+XCy,. •
The method used to determine KLA based on experimental data is presented in Section 3.6.
2-15
-------
2.3.2. Washing Machine Models
Typical operation of a residential washing machine consists of the following sequence of
events: fill, wash, spin, fill, rinse, and spin. The fill cycle consists of a falling film that impacts
an underlying pool that continuously increases in depth. Chemical volatilization may be
attributed to the falling film, splashing at the surface, and entrained air bubbles. The wash and
rinse cycles both involve agitation of the basin water for a specific length of time. The only
differences between the wash and rinse cycles are the presence of detergent for the wash cycle and
the time of agitation. The primary mass transfer mechanism associated with these cycles is
volatilization across the agitated water/air interface. Finally, during a spin cycle the washing
machine basket is rotated at a rapid rate such that the respective wash and rinse water is removed
from the clothing and pumped from the machine. It was assumed that minimal chemical
volatilization occurs during a spin cycle because of lower water volume and shorter contact time
between the contaminated water and headspace air. The rate of chemical volatilization from a
washing machine was characterized through independent investigations of the fill and wash/rinse
cycles.
2.3.2.1. Washing Machine Fill Cycle
A washing machine fill cycle is represented in Figure 2-5. Water enters a machine with a
constant inlet concentration and accumulates in the washing machine basin. On the basis of
visual observation of turbulence in the underlying pool and a short residence time for the falling
film, volatilization from the latter was assumed to be insignificant. A mass balance on the liquid
phase is based on a CFSTR such that:
(2-25)
dt
where
C, -
VI =
t ~
Q. =
Q,in =
KL =
Hc
A
chemical concentration in water (M/L3)
volume of water (L3)
time(T)
liquid fill rate (L3/T)
inlet liquid-phase concentration (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
interfacial surface area between water and adjacent air (L2).
2-16
-------
Qi,cUl
Figure 2-5. Washing machine fill cycle model.
The gas phase of the system is also assumed to approach a continuous-flow stirred-tank
reactor. A corresponding mass balance leads to:
(2-26)
where
t
Q =
HO =
A =
chemical concentration in air adjacent to water (M/L3)
headspace volume (L3)
time(T)
ventilation rate (L3/T)
gas concentration entering system from outside air (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
chemical concentration in water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas)
interfacial surface area between water and adjacent air (L2).
As with the gas-phase mass balance for dishwashers, Equation 2-26 may be simplified by
assuming that the background air is relatively clean. However, unlike the dishwasher gas-phase
mass balance, the washing machine gas-phase volume is changing with time as the liquid fills the
machine and must therefore remain as part of the derivative.
In order to determine KLA during filling, Equations 2-25 and 2-26 were solved
simultaneously. However, since the liquid and gas-phase volumes and chemical concentrations
2-17
-------
Figure 2-6. Washing machine wash/rinse cycle model.
are changing with time, a numerical solution technique was adopted to determine KLA. This
method is presented in Section 3.6.
2.3.2.2. Washing Machine Wash/Rinse Cycles
The liquid phase for a wash/rinse cycle was treated as a well-mixed batch reactor with a
constant liquid volume (Figure 2-6). The liquid- and gas-phase mass balances for this system are
identical to the liquid- and gas-phase mass balances for a dishwasher (Equations 2-20 and 2-22).
Thus, the solutions given in Equations 2-23 and 2-24 also apply for this source.
2.3.3. Shower Models
Mass balance equations for a shower (as shown in Figure 2-7) were developed previously by
Little (1992). A summary of these equations, and their derivation, is provided in the text below.
Little (1992) identified the regions within a shower system where mass transfer occurs: drop
formation, drop acceleration to terminal velocity, fall of drop at terminal velocity, and impact of
drop on shower stall surfaces. To predict the rate of mass transfer from liquid droplets to the
surrounding air, Little modeled the liquid phase as a plug-flow system with the following mass
balance:
where
Q
KL
KL C, - P
dQ_ V H.J
dz Q,
chemical concentration in water (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
(2-27)
2-18
-------
Hc
Qi
z
p
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3]iq/L3gas)
liquid flowrate (L3/T)
direction of flow
perimeter of water stream (L).
The residence time of a water droplet was assumed to be relatively short. This assumption
allowed the gas-phase concentration in Equation 2-27 to be considered constant. The resulting
differential equation was solved to give:
where
C
l.out
cun
KL
PL
P
L
1-exp -
(2-28)
outlet chemical concentration in water (M/L3)
inlet chemical concentration in water (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
interfacial area (L2)
perimeter of water stream (L)
length of stream of water (L)
Qi
Q,in
Qi
c
l,out
Figure 2-7. Shower model.
2-19
-------
Hc
liquid flowrate (L3/T)
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas).
Little (1992) also developed a mass balance to characterize the change in gas-phase
concentration during shower operation. He modeled the gas phase as a CFSTR with the transfer
of mass from the liquid being the difference between the mass flowrate of chemical entering the
system and the mass flowrate of chemical leaving the system (QA in - QiClout). The following
equation was applied:
yg
dC
-c
(2-29)
where
t
Q,
Q,
c,
chemical concentration in air adjacent to water (M/L3)
Shower stall volume (L3)
time(T)
liquid flowrate (L3/T)
inlet chemical concentration in water (M/L3)
outlet chemical concentration in water (M/L3)
ventilation rate (L3/T),
inlet chemical concentration in air (M/L3).
The solution to the gas-phase mass balance was achieved by substituting Equation 2-28 into
Equation 2-29 and then integrating, to yield:
l.out
Tt
C =—+
s D
where
Cg,o -
initial gas concentration (M/L3)
tune (T)
R
C n- — exp(- Dt)
s'° D)
(2-30)
B
(M/L3-T)
2-20
-------
Qi =
cun =
KL =
A =
Qg =
liquid flowrate (L3/T)
inlet chemical concentration in water (M/L3)
overall mass transfer coefficient for the chemical of interest (L/T)
interfacial area (L2)
ventilation rate (L3/T)
inlet chemical concentration in air (M/L3)
shower stall volume (L3)
D =
(1/T)
• Hc = Henry's law constant for chemical of interest (L liq/L gas).
Values of KLA may be estimated using Equations 2-28 and 2-30. The corresponding solution
technique is presented in Section 3.6.
2.3.4. Bathtub Models
Three distinct operations are associated with a bathtub: (1) water flowing through the faucet
with the drain open (flow-through), (2) filling the tub, and (3) bathing in a filled tub. Each
operation has a different associated mass balance model.
2.3.4.1. Bathtub Flow-Through Model
When water is flowing through the faucet with the drain open, there is no accumulation of
water in the basin. This type of operation is depicted in Figure 2-8 and may be treated in a
manner similar to that of a shower, such that Equations 2-27 and 2-29 apply.
2.3.4.2. Bathtub Fill Model
The filling of a bathtub (Figure 2-9) is similar to the washing machine fill cycle, whose mass
balance equations were given in Section 2.3.2.1. The numerical solution technique adopted for a
bathtub is provided in Section 3.6.
2.3.4.3. Bathtub Surface Volatilization Model
Finally, once the tub is filled, chemical mass transfer may continue across the water/air
interface. As shown in Figure 2-10, there are no more inputs or outputs of mass, such that the
2-21
-------
system may be modeled in a manner similar to that of a dishwasher (see Section 2.3.1) and
washing machine wash/rinse cycles (see Section 2.3.2.2).
2.4. CHEMICAL EMISSION MODELS
A valuable product of this work is the ability to predict human inhalation exposure to
contaminants present in drinking water. The level of human exposure is directly related to the
gas-phase chemical concentration, which may be estimated using the appropriate source-specific
mass balance models presented in Section 2.3. In addition to Cg, system and environmental
conditions are also important for predicting human exposure, for example, room volume, air
exchange rate, or headspace ventilation rate.
For dishwashers and washing machines, humans are exposed to chemicals emitted from the
headspace within each respective machine. This emission rate is equivalent to:
(2-31)
Qi
Q,in
a«
4
,ou
Figure 2-8. Bathtub flow-through model.
V0
Figure 2-9. Bathtub fill model.
2-22
-------
where
-"chem
Qs
a
Q
v,
Figure 2-10. Bathtub surface volatilization model.
chemical mass emission rate (M/T)
machine headspace ventilation rate (L3/T)
time-dependent gas-phase chemical concentration in machine headspace (M/L3).
Equation 2-31 may be used to estimate a chemical mass emission rate profile for the
duration of machine operation. Integration under this curve results in total chemical mass emitted
during the event. Resulting human exposure may be predicted by incorporating the mass emission
rate profile into a mass balance on the associated room air. An application of these models is
presented in Chapter 8.
In showers and bathtubs, humans receive a more "direct" exposure to volatilized chemicals.
Assuming no other losses, the mass flowrate for a plug-flow system is equivalent to:
where
E
Q, =
C
C, in =
,out
Ql (Q,in " M,out)
= chemical mass emission rate (M/T)
liquid flowrate (L3/T)
liquid-phase concentration entering system (M/L3)
= liquid-phase concentration leaving system (M/L3).
(2-32)
Mass emissions during a bathing event may be determined using:
2-23
-------
where
KL
A
H. =
(2-33)
= chemical mass emission rate (M/T)
overall mass transfer coefficient for the chemical of interest (L/T)
interfacial area (L2)
chemical concentration in water (M/L3)
chemical concentration in air adjacent to water (M/L3)
Henry's law constant for chemical of interest (L3liq/L3gas).
As with dishwashers and washing machines, the mass emission rate profile may be
developed for the duration of a showering or bathing event. Example applications of mass
emission models are presented in Chapter 8.
2-24
-------
3. GENERAL METHODOLOGY
There is currently no standardized protocol for measuring volatilization rates from indoor
sources. _Thus, a general protocol was developed for this project and was applied to all sources.
Common features of this protocol follow:
1. A stainless steel source chamber
2. Chemical tracers
3. Chemical sampling
4. Sample analysis
5. Quality assurance measures
6. Data analysis .
7. Mass closure assessments
The experimental methodology and associated quality assurance measures applicable to all
sources are presented in this section. Methodologies specific to individual sources are discussed
in the respective source chapters (Chapters 4 to 7).
3.1. SOURCE CHAMBER
Previous experimental studies vary in their isolation of household tap water sources of
volatilization, which range from abandoned houses to laboratories and exposure chambers. All
experiments for this project were completed within a stainless steel exposure chamber. The
chamber was 2.4 m x 1.8 m x 2.4 m (11 m3 in volume). It was ventilated under negative
pressure; that is, air was drawn into the chamber and exhausted through a ceiling port connected
to a fume hood. The enclosed chamber had the advantages of allowing for a mass closure
assessment through the measurement of both liquid- and gas-phase concentrations and
volumetric flowrates. Although washing machines and dishwashers effectively served as their
own exposure chambers, the experiments involving these sources were also completed in the
stainless steel chamber.
3-1
-------
3.2. CHEMICAL TRACERS
3.2.1. Physicochemical Properties
All experiments were completed using a chemical cocktail containing five volatile tracers:
acetone, ethyl acetate, toluene, ethylbenzene, and cyclohexane. These five chemicals represent a
wide range of Henry's law constants. Physicochemical properties for each chemical are given in
Table 3-1. The five chemicals were chosen to meet the following requirements:
8. At least one chemical had an Hc > 1.0 m3liq/m3gas at 25°C.
9. At least one chemical had an Hc < 0.005 m3Iiq/m3gas at 25°C.
10. Two chemicals had similar physicochemical properties, that is, similar Hc and molecular
diffusion coefficients.
11. All chemicals had the capability of being analyzed with the same gas chromatography (GC)
system and flame ionization detector (FID), with adequate separation of peaks.
12. Chemicals were easily identified and quantified by GC/FID at low aqueous-phase
concentrations (< 500jo,g/L) to minimize chemical usage and discharge during experiments.
13. Chemicals had a solubility > 10 mg/L in water.
14. At desired concentrations, chemicals posed minimum risks to researchers during
experiments.
Henry's law constants for chemicals used in experiments completed at temperatures other
than 25°C were adjusted to reflect the temperature change. To determine the change in Henry's
law constant with increasing or decreasing temperature, existing equations developed by
Ashworth et al. (1988) were used for toluene, ethylbenzene, and cyclohexane.
Table 3-1. Summary of physicochemical properties for selected chemical tracers
Compound
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Hc @ 25°C
(m3);n/m3Bas)
0.0015
0.0050
0.27
0.33
7.2
D, @ 24°C
(cm2/s)
1.1E-05
9.5E-06
9.1E-06
8.4E-06
9.0E-06
Dg@24°C
(cm2/s)
0.11
0.092
0.085
0.077
0.088
Tb
(°C)
56.5
77.0
110.6
136.2
80.7
P
(kg/L)
0.79
0.89
0.87
0.87
0.77
Solubility
(mg/L)
miscible
64000
515
152
58
Pv°
(mm Hg)
270
115
22.0
7.0
77
Sources: Ashworth (1988), CRC Handbook (1995), Howard (1990) and Tucker and Nelken (1990).
3-2
-------
These equations have been validated for a temperature range of 10°C to 30°C.
Toluene: HcT = exp[5.133 - 3024/(T + 273.15)]/(0.000082*(T + 273.15)) (3.1)
Ethylbenzene: HC>T = exp[l 1.92 - 4994/(T + 273.15)]/(0.000082*(T + 273.15)) (3.2)
Cyclohexane: HcT = exp[9.141 - 3238/(T + 273.15)]/(0.000082*(T + 273.15)) (3.3)
where
HcT = Henry's law constant at experimental temperature (L3Iiq/L3gas)
T = experimental temperature (°C).
Schoene and Steinhanses (1985) developed the following relationship between Henry's law
constant and temperature for acetone.
Acetone: log(HCjT) = - 2218/(T +273.15)+ 4.545 (3.4)
where
HcT = Henry'slaw constant at experimental temperature (L3Iiq/L3gas)
T = experimental temperature (°C).
There is a lack of published information related to temperature effects on Henry's law constant
for ethyl acetate. The following relationship was used to predict the change in ethyl acetate's
Henry's law constant at different experimental temperatures (Enviromega, 1993).
where
H(
H
c,T
o,25°C
Ethyl Acetate: HcT = HC)25oc« 1.044(T-25°C)
Henry's law constant at experimental temperature (L3,iq/L3g
Henry's law constant at 25°C (L3liq/L3gJ
Experimental temperature (°C).
(3.5)
3.2.2. Chemical Tracer Addition
The water used in each experimental system was spiked with a multitracer stock solution.
Tracer solutions were prepared in 3 L Tedlar™ bags fitted with a stainless steel hose/valve with
3-3
-------
locking screw and a replaceable Teflon™-lined septum with a stainless steel cap. Tedlar™ bags were
ideal for preparing and transferring volatile tracer solutions because of the minimal headspace
associated with filling and emptying the bag with liquid. A bag had the capability to expand or
collapse without forming a headspace. Each bag was filled with cold tap water using a variable-
speed peristaltic pump (Masterflex™ Laboratory Standard Variable Speed Drive System). Teflon™
tubing (0.635 cm OD) dedicated to clean water usage provided the means of water transfer through
the inlet valve of each bag. Any air added during this procedure was removed by collecting it in a
large bubble near the bag's valve opening and reversing the pump, thereby emptying the excess air.
Syringes adequately cleaned with methanol and water were used for chemical injections. Known
amounts of each chemical were injected into a known volume of tap water contained in the Tedlar™
bag. In addition to resulting system concentrations, chemical solubilities were considered when
detenriinmg the volume of chemical to be added. To facilitate dissolution, bags were manually
agitated and allowed to sit for periods over 24 hours prior to use in any experiments.
The predissolved solution had to be added to the water supply of each experimental system. To
add the chemical solution, the Masterflex™ peristaltic pump described above with Teflon™ tubing
dedicated to chemical addition was used. As the bags were emptied, the chemical solution was
manually mixed into the system's water supply. Additional experimental procedures are presented in
each respective source chapter (Chapters 4 to 7).
3.3. CHEMICAL SAMPLING
3.3.1. Liquid-Phase Sampling
Each experimental system was retrofitted with a liquid sample port made of Teflon™. This port
was designed to miiiimize chemical losses during sampling, for example, preventing air in the
sample line. In addition, Teflon™ tubing was connected to the end of the sample port such that when
a liquid sample was collected the water entered near the bottom of the sample vial, thereby
mhiimizing splashing. Liquid samples were collected in 22 mL glass vials and sealed with an
aluminum cap fitted with a Teflon™-faced silicon septum. Approximately 11 mL of water were
collected in each vial, leaving a significant headspace in the vial.
3-4
-------
Samples were stored at 4°C in a laboratory refrigerator until analysis. Preexperimental tests
determined that samples could be stored up to 1 week at 4°C without significant losses. Samples
moved to another location for analysis were transported in an ice chest at a temperature at or below
4°C. A data log book contained a record of each liquid sample including information regarding date
of collection, date of transport (if necessary), length of storage, and date of analysis.
3.3.2. Gas-Phase Sampling
Each experimental system was also retrofitted with a stainless steel Swagelok™ gas sample port.
Gas samples were collected on Carbotrap™ 300 (Supelco™) adsorbent tubes (0.635 cm OD x 17.8
cm). Carbotrap™ adsorbent tubes were packed with graphitized carbon black and were determined
to be suitable for trapping and thermally desorbing the target organic compounds. Samples were
collected using a gas sample pump and bubble flowmeter in series as shown in Figure 3-1. The
sorbent tube was attached to the Swagelok™ port open to the experimental system's headspace from
which gas was drawn. The flowrate at which gas was drawn through the sorbent tube was measured
by the bubble flowmeter. The volume of air drawn through the sorbent tube was determined by
timing the event. Sample flowrates were in the range of 0.15 to 0.55 L/minute. For batch and
flowthrough experiments, sampling times ranged from 30 to 60 seconds and were scheduled such
that a liquid sample was collected during the gas sampling period. For fill cycle experiments, a
, single'gas sample was collected for the duration of the 3- to 6-minute experiments. Preliminary tests
were completed to ensure that experimental sample tubes were not achieving breakthrough at these
sampling conditions.
Once the gas sample had been collected, the ends of the sorbent tube were sealed with stainless
steel Swagelok™ caps and stored at 4°C in a hermetically sealed jar containing activated carbon.
Again, it was determined through preexperimental testing that gas samples could be stored up to 1
week without sample loss at these conditions.
3-5
-------
Sorbent Tube
Swagelok™ Fitting
" V ,
ibe U- — 1
~
V4
^
Bubble Flowmeter
Gas Sample Pump
Figure 3-1. Gas sampling experimental setup.
3.4. SAMPLE ANALYSES .. . .
3.4.1. Liquid Sample Analysis
Liquid samples were analyzed using a headspace concentrator equipped with an autosampler
(Tekmar 7000) and a gas chromatograph (Hewlett Packard, 5890 Series II Plus) with a flame
ionization detector (GC/FID). Method parameters for the headspace concentrator were based on
previous experimental work and were as follows: sealed vials containing liquid samples were
lowered into a platen chamber where they were heated at 70°C for 60 minutes, allowing the volatile
organic compounds to be transferred into the vial headspace and to reach equilibrium. Equilibrium
concentration is highly dependent on temperature, and a platen temperature of 70°C was determined
to be an optimum value for this study. Each liquid sample was heated for identical periods of time
and temperatures, thus enhancing reproducibility. Following the platen equilibration time, the vial
was pressurized with helium for 1 minute. The sample loop was then filled for 1 minute and allowed
to stabilize for 0.2 minutes. To prevent condensation, the sample loop temperature was set at 100°C.
The headspace sample was then injected for one minute into the gas chromatograph at a temperature
of!00°C.
The GC/FID parameters included an inlet temperature of 200°C and a detector temperature of
250°C, once again preventing condensation. For each sample, the initial oven temperature was
32°C, which was held constant for 0.5 minutes before being ramped at 20°C/minute to a final oven
temperature of 55°C. This final temperature was held constant for 1 minute, leading to a total run
time of 2.65 minutes. Over the course of the experimental period, different GC/FID columns were
3-6
-------
used. Analytical columns included a Restek™ capillary column (30 m-x 0.53 mm x 3.0 jam film
thickness) and an HP-1 capillary column (5 m x- 0.53 mm x 2.65 |j,m film thickness).
Chromatographic peaks were drawn and integrated using HP 3365 Series II ChemStation (Version
A.03.34) software. A sample chromatogram is shown in Figure 3-2, where the abscissa is time and
the ordinate is the GC/FID response.
Liquid samples were analyzed within 1 week of collection, typically on the same day of
collection. To limit column contamination, the vials were placed on the headspace autosampler tray
in the order of increasing concentration. Vials containing clean water (blanks) were placed
intermittently between sample vials, and served as indicators of system contamination, which was
always minimal.
3.4.2. Liquid Standards
For each experiment, an additional stock solution was prepared as described in Section 3.2.2 for
the preparation of liquid standards. Ten milliliters of cold tap water were added to a 22 mL glass
vial using a volumetric glass pipette. The vial was sealed with a Teflon™-faced silicon septum and
aluminum cap using a hand-held crimper. A small volume of stock solution with a known
concentration of each chemical was injected into a solution with a known volume of clean water.
3.0e5
2.0e5 -
1.0e5
o -
0.0
Toluene
Ethylbenzene
Ethyl Acetate 1 1
Acetone
. A Cyclohexane M
A A ,A , .11 : , A .
0.5 1.0 1.5 2.0 2.5
Time (min)
Figure 3-2. Liquid-phase sample chromatogram.
3-7
-------
The resulting concentrations of each chemical in the vial were calculated. By varying the amount of
stock solution extracted from the bag, six vials with different concentrations were prepared. These
standards were used to develop a six-point linear calibration curve, for example, gas chromatograph
area response versus chemical concentration. The six calibration points were chosen based on
experimental data, such that liquid sample measurements were within the range of standards. A
sample liquid calibration curve is shown in Figure 3-3. External calibration curves for each tracer
had a coefficient of determination (R2) of at least 0.95 and were nearly always greater than 0.98.
Since several experiments incorporated the use of detergent, a separate test was completed to
determine the effects of detergent on chemical calibration curves. Using a single stock solution, two
calibration curves were developed: one using clean (no detergent present) water as the matrix and a
second using "soapy" water as the matrix. In a comparison of the two curves, only ethyl acetate was
significantly affected by the presence of detergent in the water. This impact is further discussed in
appropriate source-specific chapters.
As described in Section 3.2.2, it is conceivable that complete dissolution of some tracers,
particularly cyclohexane, did not occur prior to the development of calibration curves. In such cases,
liquid-phase concentrations may have been overestimated. This is likely the reason for relatively
poor mass closures for cyclohexane during many experiments. However, stripping efficiencies and
mass transfer coefficients for cyclohexane should not have been affected by this phenomenon. This
3000000
2500000--
I 2000000--
o
§* 1500000--
O 1000000--
500000--
0
0
y=126200x
R2 = 0.9992
10 15
Liquid Concentration (mg/L)
20
25
Figure 3-3. Liquid-phase calibration curve for ethylbenzene.
3-8
-------
is true because all stripping efficiencies and mass transfer coefficients for cyclohexane were based on
relative changes in liquid concentrations.
3.4.3. Gas Sample Analysis
Gas samples were analyzed using a thermal desorber with an autosampler (Tekmar 6016) and a
purge and trap concentrator (Tekmar 3000). Over the course of the experimental period, this system
was plumbed to different gas chromatographs. First, it was plumbed to the GC/FID described hi
Section 3.4.1. Most recently, the system was plumbed to a gas chromatograph (Hewlett Packard,
6890 Series) with a flame ionization detector (GC/FID#2). Method parameters for the thermal
desorber and purge and trap system were based on recommended values for Carbotrap™ 300 sorbent
tubes. Each tube was heated at 200°C for 8 minutes. The desorbed contaminants were transported
to the purge and trap column through a transfer line with a temperature of 200°C. Once the
desorption phase was complete, the trap was heated to 250°G for 2 minutes. During this time,
contaminants were desorbed from the trap and immediately injected into the GC/FID#2.
The GC/FID#2 method for gas samples included an inlet temperature of 225°C and a detection
temperature of250°C, once again preventing condensation. For each sample, the initial oven
temperature was 34°C, which was held constant for 0.5 minutes before being ramped at 10°C/minute
to a final oven temperature of 65°C. This final temperature was held constant for 11 minutes,
yielding a total run time of 14.6 minutes. The primary analytical column for GC/FID#2 was a
Restek™ capillary column (30 m x 0.53 mm x 3.0 |j,m film thickness). Chromatographic peaks were
drawn and integrated using HP GC ChemStation (Version Rev. A.04.02) software. A sample
chromatogram for GC/FID#2 is shown in Figure 3-4, where the abscissa is time and the ordinate is
the GC/FID response.
3.4.4. Gas Standards
A pressurized gas cylinder of known concentration of each chemical tracer was purchased
(calibrated by Scott Specialty Gases, NIST traceable to Project 0454764). The cylinder contained a
balance gas of air and was certified to contain: 40.0 ppm acetone, 50.6 ppm ethyl acetate, 40.5 ppm
3-9
-------
PA-
1750 -:
1500 i
1250 -I
1000 -i
750 -i
500 •:
250 -
o-i
1™. i A x Toluene
Ethyl Acetate
Ac«
i ir-.i/vi
;tone
Cyclohexane
i
Ethylt
-
, j
tenzene
| • • • | • • • I . . . I . • . . , . . . , . I I | i
2 4 6 8 10 12 14
Time (min)
Figure 3-4. Gas sample chromatogram.
toluene, 27.7 ppm ethylbenzene, and 19.9 ppm cyclohexane. A 3 L Tedlar™ sample bag dedicated to
gas standards was filled with the gas stock solution from the tank. A sampling configuration similar
to the one shown in Figure 3-1 was used to draw the standard gas from the Tedlar™ bag through a
clean sorbent tube. The volume of gas drawn through the tube was measured using a bubble
flowmeter and a stopwatch. Different gas volumes were drawn through five sorbent tubes resulting
in a five-point external calibration curve as shown in Figure 3-5. As with liquid standards, gas
standards were prepared for each experiment in accordance with expected gas-phase measurements.
External calibration curves for each tracer had a coefficient of determination (R2) of at least 0.999.
200000
150000 -
100000-
50000 -
-i :—i 1 r i r
0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045
Mass
Figure 3-5. Gas-phase calibration curve for acetone.
3-10
-------
3.5. QUALITY ASSURANCE MEASURES
A quality assurance plan was developed specifically for this project and was submitted to the
US Environmental Protection Agency at an earlier date (September 1996). This plan was
implemented throughout the entire study. A summary of quality assurance measures is given in this
section.
3.5.1. Duplicate Samples
Because of the high volatility associated with several of the chemical tracers, duplicate liquid-
phase samples were collected for every experiment. For the purposes of this study, duplicate
samples refer to samples that were collected sequentially and that differed in time by fewer than 20
seconds. A summary of results associated with duplicate samples is presented in Table 3-2. The
average difference reported in Table 3-2 includes all duplicates, even those that were removed for
violating the quality assurance project plan.
The best duplication for liquid samples was achieved for ethyl acetate, with an average relative
difference of 2.5%. Only 7.9% of all liquid sample duplicates had differences of greater than 20%.
Twenty-six of the 38 liquid samples with poor duplication (>20% difference) were not included in
• " *'?
the data analysis used to predict volatilization parameters. The remaining 12 duplicates had a
relative difference between 22% and 36%, and were collected during the initial seconds of
dishwasher experiments. As is explained in Chapter 5, chemicals rapidly volatilized from water used
in a dishwasher within the first 45 seconds of operation. To characterize this drop in liquid-phase
Table 3-2. Duplicate sample results
Liquid Samples
Compound
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Number of Duplicates
113
67
111
101
96
Average Difference3 (%)
3.1
2.5
7.7
8.2
13
"Defined as -V-S-rS-.lOO.
3-11
-------
concentration, an initial value was needed. Dishwasher experiments using the average of these
duplicates as initial concentrations are flagged in Chapter 5.
3.5.2. Replicate Experiments
Replicate experiments refer to experiments completed under approximately identical conditions,
but not sequentially on the same day. For this study, 25% to 38% of experiments for each system
were repeated. A summary of replicate experimental results is provided in Table 3-3.
Table 3-3 includes replicate experimental results for all sources. In fact, some replicate
experiments for bathtubs and washing machine fill cycles were particularly poor for reasons that are
explained later in this report. If those experiments are excluded from the analysis, the replicate
sample results are improved significantly as presented in Table 3-4. The best replicate sample results
were achieved for toluene, as shown in Table 3-4.
Table 3-3. Replicate sample results
Volatilization Parameters
Compound
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Average
Difference3
forti(%)
38
23
16
18
20
Median
Difference3
fom(%)
30
26
6.4
6.9
5.0
Average
Difference1"
forKLA(%)
26
26
23
22
28
Median
Difference"1
forKLA(%)
19
27
19
17
19
"Defined as —]
10°-
'Defined as i'.100.
3-12
-------
Table 3-4. Replicate sample results excluding replicate experiments associated with filling
Volatilization Parameters
Compound
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Average
Difference3
for n (%)
22
18
6.1
7.0
7.9
Median
Difference8
for t| (%)
16
17
4.2
3.4
1.8
Average
Difference15
forKLA(%)
27
29
17
13
18
Median
Difference15
forKtA(%)
22
28
12
9.9
16
defined as-
• 100.
Defined as E
n "^
3.5.3. Experimental Blanks
A minimum of four analysis blanks were analyzed for every batch of experimental samples.
These blanks were prepared in the laboratory and treated as an experimental sample through the
analysis phase. The concentration of each volatile tracer hi each blank was always below detection
limit.
3.5.4. Method Detection Limit
The method detection limit was defined as:
where
MDL
MDL=t(n-U-«=0.95)*Sr
(3.6)
, method detection limit
standard deviation of replicate analyses (M or M/L3)
= student's t value for a one-sided 95% confidence level and a standard
deviation estimate (sr) with n-1 degrees of freedom.
3-13
-------
The MDLs were determined separately for liquid and gas samples. Results are listed in Table 3-5. It
should be noted that the MDL test was completed for each GC/FID and column combination used in
liquid and gas sample analyses. A majority of the liquid samples was analyzed using the 5 m column
in GC/FID #1, and a majority of the gas samples was analyzed using the 30 m column in GC/FID #2.
Less mass was used for each respective chemical to determine the MDLs for the 5 m column and
GC/FID #2 than was used to determine the MDLs for the 30 m column and GC/FID #1.
Subsequently, the standard deviation was lower for each chemical (sr), resulting in a significantly
lower associated MDL. If a lower mass had been used for the test with the 30m column and
GC/FID #2, the resulting MDLs would have been lower.
3.6. DATA ANALYSIS
Experimental systems with similar liquid flow patterns shared the same data analysis methods.
The sources were grouped as follows:
1. Batch systems: dishwasher, washing machine (wash/rinse cycles), and bathtub (surface
volatilization)
2. Plug-flow systems: shower and bathtub (flow-through)
3. Fill systems: washing machine (fill cycle) and bathtub (fill process)
The methods used to predict chemical stripping efficiencies, KLA, kg/k,, and k, and kg for each
type of experimental system are described below. The procedures in this section were applied
independently for each chemical tracer. Thus, unless otherwise stated, five separate values were
Table 3-5. Method detection limits (MDLs) for liquid and gas samples
Chemical
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Liquid MDL
for 30 m column3
(mg/L)
0.80
0.46
0.23
0.33
0.16
Liquid MDL
for 5 m column9
(mg/L)
0.12
0.13
0.09
0.09
0.07
Gas MDL
for GC/FID #lb
(US)
1.9
6.1
4.0
6.8
0.22
Gas MDL
For GC/FID #2b
(US)
0.42
1.3
1.2
1.2
0.30
'Both 30m and 5 m columns were used in GC/FID #1 for all liquid samples.
"The same 30m column was used in both GC/FID #1 and GC/FID #2 for all gas samples.
3-14
-------
reported for each volatilization parameter. Deviations from the solution techniques in this section
are discussed in Chapters 4 to 7.
3.6.1. Chemical Stripping Efficiencies
Equations 2.1 and 2.2 were used to determine chemical stripping efficiencies for all
experiments. For batch and fill systems, Equation 2.2 was used with C, end equal to the final liquid-
phase chemical concentration measurement, and C, init equal to the measured liquid-phase chemical
concentration before starting the experiment.
Equation 2.1 was used for the plug-flow systems, where CIin was equal to the liquid-phase
chemical concentration in the tracer reservoir and Q out was equal to the liquid-phase chemical
concentration in the specific system at the drain. When chemical volatilization in the tracer reservoir
was a concern, that is, the inlet chemical concentration was changing, chemical stripping efficiencies
were determined for several periods of the experiment. Each period consisted of at least one
reservoir liquid-phase measurement and at least one system liquid-phase and gas-phase
measurement The stripping efficiency reported for the experiment was an average value based on
each period;
3.6.2. Overall Mass Transfer Coefficients (KLA)
In Section 2.3, mass balance models were developed for each experimental system. These
models served as a way to determine KLA for each chemical tracer and source operating condition.
Most of the models could not be solved analytically to determine KLA. Thus, an iterative solution
technique was adopted.
For batch systems, liquid-phase and gas-phase concentrations were predicted using Equations
2.23 and 2.24, respectively, for a given KLA value. For each experiment, liquid- and gas-phase •
chemical concentrations were measured at a given time. To determine the best KLA associated with
these measurements, the mathematical models represented by Equations 2.23 and 2.24 were "fitted"
with the best KLA value by minimizing the sum of squared normalized residuals between modeled
and measured concentrations:
3-15
-------
Sum of square of residuals = Y1
T- Z_i
C — C
(3.7)
where
-"expt
experimentally measured liquid and gas concentrations (M/L3)
mathematically predicted liquid and gas concentrations (M/L3).
Equation 3.7 was minimized using two different approaches: (1) based on liquid-phase
measurements only and (2) based on gas-phase measurements only.
An Excel™ Spreadsheet solver was used to complete the iterations. Ideally, the two best-fit
values of KLA should be the same. However, as is explained in a later section, this often was not the
case. Gas samples were collected for a longer sampling time than liquid samples. Because liquid
samples were more representative of actual conditions at specific experimental times, they were used
to predict values of KLA. However, in some cases, the change in liquid-phase concentration for
acetone and ethyl acetate was relatively small for the duration of an experiment. Thus, a general
protocol was developed such that when overall stripping efficiencies for a given chemical
approached the value of error associated with duplicate samples (see Table 3-2), associated values of
KLA were based solely on gas-phase data. The solution techniques for each chemical are described
in Chapters 4 to 7.
For fill systems, the differential mass balance equations (Equations 2.25 and 2.26) could not be
solved analytically. Thus, a second-order Runge-Kutta numerical solution technique was used with
1-second tune steps to determine the liquid- and gas-phase concentrations for a given value of KLA.
The second-order approximations of Equations 2.25 and 2.26 are predicted with the following
equations:
C/«*> =C/" + -
+1 =cg" -
(3.8)
(3.9)
3-16
-------
where
t n+l _
= chemical liquid-phase concentration at time step n+l (M/L3)
C," = chemical liquid-phase concentration at time step n (M/L3)
At = differential time step (T)
f(tn,C,n)
c
Q.C..i
..in
wn
QiC,11 _ K^AC," + KLACg
v,n v," V;HO
ijin = inlet liquid-phase chemical concentration (M/L3)
Qi = liquid fill flowrate (L3/T)
V,n = ^liquid volume at time step n (L3) = Q,«t
t = time(T)
KL = overall mass transfer coefficient (L/T)
A = interfacial surface area between water and adjacent air (L2)
Hc - Henry's law constant (L3liq/L3gas)
Cgn+1 = chemical gas-phase concentration at time step n+l (M/L3)
Cgn = chemical gas-phase concentration at time step n (M/L3)
f(tn,Cg") =
QA" + Q.c8n + KLAC,n _ KLACgn
.(Vt-V,n) (Vt-V,n) (V,-V,B) (vt.-V,')He
For plug-flow systems, KLA was determined by using Equation 3.7 with measured data and
concentrations predicted by Equation 2.28 or Equation 2.30. Values of KLA were determined for
different experimental periods and then averaged.
3.6.3. Ratio of Gas-to-Liquid Phase Mass Transfer Coefficients
An important component of this study involved the determination of kg/kj ratios for each
experiment. Previous research has shown this ratio not to change significantly between chemicals
for a given experimental system and operating conditions (Munz and Roberts, 1989). Thus, a single
k^ value was estimated for each experiment, which was assumed to be constant for all chemicals.
To determine this kg/k, value, the following steps were followed:
3-17
-------
For a given experiment, the value of KLA for each chemical was determined as outlined in
Section 3.6.2.
Using the experimentally determined values of KLA, the ratio of KLAj/KLAj for all combinations
of chemicals (^AmaaJKiAtaiylKeMe, 'K~LAacetone/KL\ll^ne, etc.) was calculated and organized in a
5x5 matrix (Matrix 1) as shown in Figure 3-6.
The ratio of KLA;/KLAj for each chemical combination was also predicted using Equation 2.15
Q¥m~) and a single assumed value of kg/k,. These predicted ratios were organized in a second 5 x
5 matrix (Matrix 2) also following the format of Figure 3-6.
Equation 3.7 was used to calculate the normalized residuals between the measured ratios of
Matrix 1 and the predicted ratios of Matrix 2. These residuals were placed in the associated
column and row position in a third 5x5 matrix (Matrix 3).
All of the entries in Matrix 3 were summed to find the total residual between Matrix 1 and
Matrix 2. The total residual was minimized by choosing different values of kg/k, used to predict
KLA values hi Matrix 2. The value of kg/k, which led to a minimum total residual between
measured and predicted values was recorded and used for a given experiment.
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
Acetone
Ethyl
Acetate
Toluene Ethylbenzene Cyclohexane
^-L-"-ace 1
KLAace
KLAea
KiA^
KiAt0i
KLA..
K^
KLAace
K-LAcyd,,
KLA^
KLAace
KLAea
KLAea -j
KLA.
KLAto,
KLA.
KLAeb
K^Aea
1
-------
3.6.4. Liquid- and Gas-Phase Mass Transfer Coefficients
Once the KLA values for each chemical and a single k^/k, value were determined for an
experiment, liquid- and gas-phase mass transfer coefficients were calculated for each chemical. The
overall resistance equation (Equation 2.5) was used to solve for k,A, where kgA was written in terms
of k,A using the kg/k, value. Once k,A was determined, the value for kgA was predicted. Although,
was constant for all chemical tracers, k,A and kgA were compound dependent.
3.7. FACTORIAL ANALYSIS
A factorial analysis was used to determine the main effects associated with the primary
experimental variables (Box and Bisgaard, 1988). For dishwasher, washing machine wash/rinse
cycle, and shower experiments, 2x2x2 factorial arrays were designed. For these arrays, the main
effect for a single variable was calculated as the average of the difference between responses at two
levels of the factor of interest. Variable responses were in terms of stripping efficiencies and KLA
values. This procedure was completed for all three factorial variables. The largest positive or
negative value corresponded to the largest main effect.
3.8. MASS CLOSURE ASSESSMENT
Previous research related to the volatilization of chemicals from drinking water, in particular
showers, has suffered from poor mass closure and, in some cases, lack of adequate experimental
measurements to assess mass closure. Therefore, an important protocol for each source experiment
was to obtain adequate mass closure. Mass closure for batch systems was determined using:
where
V =
cu =
% mass recovered =
V,Cu+VgCg,2+QgJCgdt
v,cu+vgc&1
liquid volume (L3)
headspace volume (L3)
chemical concentration in liquid phase at tune 1 (M/L3)
chemical concentration in liquid phase at time 2 (M/L3)
3-19
(3.10)
-------
Cg,2 =
Qs =
t, =
chemical concentration in gas phase at time 1 (M/L3)
chemical concentration in gas phase at time 2 (M/L3)
ventilation rate of system (L3/T)
time 1 (T)
time 2 (T).
For fill systems, mass closure was determined by:
% mass recovered =
VA,2+VgCg,2+QjCgdt
Q.CuA-O+v^
where
V, -
c,.in =
Ce,i =
Qi -
Qg =
t, =
liquid volume (L3)
headspace volume (L3)
inlet chemical concentration (M/L3)
chemical concentration in liquid phase at time 2 (M/L3)
chemical concentration in gas phase at time 1 (M/L3)
chemical concentration in gas phase at time 2 (M/L3)
liquid flowrate in and out of system (L3/T)
= ventilation rate of system (L3/T)
= time 1 (T)
= time 2 (T).
(3.11)
Finally, for plug-flow systems, the following mass closure equation was used:
% mass recovered =
(3.12)
where
VB = headspace volume (L3)
.
3-20
-------
Ci,m = average chemical concentration measured in tracer reservoir (M/L3).
C,out= chemical concentration in liquid phase at outlet (M/L3)
Cg = chemical concentration in gas phase (M/L3)
Cg j = chemical concentration in gas phase at time 1 (M/L3)
Cg 2 = chemical concentration in gas phase at time 2 (M/L3)
Q! = liquid flowrate in and out of system (L3/T)
Qg = ventilation rate of system (L3/T)
t, = time 1 (T)
tj = time 2 (T).
For most experimental systems, Cg was measured at several times during an experiment. In
these cases, there were periods where the exact gas concentration in the system's headspace was not
known. To account for these unknown values in Equations 3.10 to 3.12, the concentrations of
samples collected on either side of this period were averaged.
Mass closure results for all chemicals are reported in each respective source section. Adequate
mass closure was defined as 75% to 125%.
3-21
-------
-------
4. SHOWER EXPERIMENTS
Shower operation consists of a single water activity, that is, no separate cycles. To study
this activity, a wide range of operating conditions were applied to a consistent experimental
design.
4.1. EXPERIMENTAL SYSTEM
A 140 cm x 70 cm x 178 cm (1.7 m3 total volume) shower stall (with bathtub) was
purchased to complete all shower experiments. The shower stall was installed in the stainless
steel chamber on a 58 cm high cinder-block platform. The platform served two purposes: (1) it
elevated the system to an appropriate height for draining and collecting liquid samples and (2) it
elevated the shower stall such that it reached the stainless steel chamber's ceiling, which
provided a system boundary. Other system boundaries included three walls and a floor made of
fiberglass coated with an unknown plastic, and one wall (a curtain) made of Tedlar™.
Showering involves production of a spray of water that impacts on and cascades down
surfaces to the bathtub floor. The floor slopes toward a drain where water is removed from the
system. The experimental shower system required an auxiliary water supply (see Figure 4-1).
To meet this need, the washing machine described in Section 6.1.1 effectively served as a tracer
reservoir. The washing machine was directly plumbed to the building cold and hot water supply.
Chemicals were added to the washing machine as it filled (~ 90 L). The reservoir's contents
were further mixed by using wash cycle agitation. The washing machine contents were pumped
with a rotary vane pump (PROCON™) through 1.3 cm OD Teflon™ tubing to the shower head.
An adjustable low-flow (9.5 L/minute maximum) showerhead (Interbath™) was used for all
experiments. The showerhead could be adjusted between fine and coarse spray. A 60 mm, 19
L/min maximum rotameter (King Instrument Co.) was installed hi the Teflon™ tubing line to
measure the liquid flowrate through the system. The experimental flowrates were based on
typical values and the restrictions of the showerhead. The accuracy of the rotameter was verified
by timing the collection of a known volume of liquid from the showerhead.
4-1
-------
Liquid Tracer Reservoir
Temp
Probe
Cinder Block Platform
Liquid Sample
To Drain port
Figure 4-1. Shower experimental system.
Liquid samples were collected from the washing machine reservoir in a manner similar to
the actual washing machine experiments (see Section 6.1.1). The shower stall was designed to
collect the necessary samples to solve the shower mass balance equations (Equations 2-28 and 2-
30). A liquid sample port was installed in the base of the bathtub near the drain. A 30 cm length
of 0.64 cm OD Teflon™ tubing with a Teflon™ sample valve was connected to this port. Liquid
samples were collected as described in Section 3.3.1.
Three gas sample ports were installed in the system to better understand the gas-phase
chemical concentration distribution in the stall. Sample port #1 was located within the chamber
exhaust vent and consisted of a 91-cm-long 0.64 cm OD Teflon™ tube attached to a stainless
steel Swagelok™ union at which point a sorbent tube was connected. Port #2 was a bore-through
Swagelok™ fitting located on the wall with the showerhead, 53 cm from the bathtub floor. Port
#3 was located on the shower curtain, 61 cm from the floor of the bathtub. A Swagelok™ fitting
was inserted in the curtain for sample collection. Because of time constraints, only gas samples
collected from sample ports #1 and #3 were collected as described in Section 3.3.2. The
4-2
-------
sampling flowrates for sorbent tubes used for sample collection at port #2 were measured and
recorded before the start of each experiment with clean air. Thus, a bubble fiowmeter was not
used hi the sampling train (see Figure 3-1) at this port.
A liquid temperature probe was submerged in the tracer reservoir, and a second probe was
inserted in the shower stall near the drain. Liquid temperatures at these two locations were
continuously measured using a thermocouple and digital monitor. The temperature difference
between these two points was minimal for all experiments.
4.2. EXPERIMENTAL DESIGN
The following operating variables were selected for shower experiments: water
temperature, liquid flowrate, and shower spray type. The impact of these operating conditions
on chemical volatilization rates was studied using a 2 x 2 x 2 factorial array. As shown hi
Figure 4-2, variable ranges were cold (T « 22°C) versus warm (T « 35°C), low liquid flowrate
(6.1 L/minute) versus high liquid flowrate (9.1 L/minute), and fine shower spray versus coarse
shower spray. Eight experiments were completed, with two additional experiments serving as
replicates.
High -
Flowrate
Low —
Flowrate
— Water Temperature-35 C
— Water Temperature ~ 22 C
Coarse
Spray
Fine
Spray
Figure 4-2. Shower factorial experimental design.
4-3
-------
4.3. SOURCE-SPECIFIC METHODOLOGY
Prior to each shower experiment, the following tasks were completed:
• Flowrates for sorbent tubes used at port #2 were measured with clean air
• The desired experimental liquid flowrate was set using the rotameter
• The tracer reservoir (washing machine) was filled with either cold or warm tap water
• The chemical tracer solution was pumped into the washing machine as it filled
• The washing machine reservoir solution was mixed by allowing the washing machine to
agitate for approximately 1 minute
• An initial gas-phase sample was collected from sample port #1 in the shower stall
• Two initial reservoir liquid-phase samples were collected.
4.3.1. Sample Schedule
Shower experiments lasted 8 minutes, during which time liquid-phase samples were
collected from both the tracer reservoir and the shower stall. Five shower stall liquid samples
were collected at experimental times of 0.5,1.5, two at 3.75, and 7.75 minutes. Although the
tracer reservoir chemical concentrations did not change significantly for most experiments, three
tracer reservoir samples and one duplicate sample were collected and scheduled within 45
seconds of each shower stall sample so that several independent stripping efficiencies could be
determined for a single experiment.
A total of 12 gas samples were collected for every shower experiment. Six gas samples
were collected at port #1 for 30 seconds and were scheduled such that a shower stall liquid
sample was collected at the midpoint of the gas sample time. Three gas samples were collected
at each port #2 and port #3. The sampling tunes at these ports were scheduled to occur
simultaneously, as well as at the same time as a gas sample collected at port #1. Thus, the gas-
phase chemical concentration distribution was determined for three separate time periods in an
experiment. Finally, a gas sample was collected after the experiment had ended and no water
flowed through the system. The start time of this sample ranged from 5 to 20 minutes after the
completion of an experiment. The gas collection time was 5 minutes.
4-4
-------
4.3.2. Ventilation Rate
Through use of a smoke test, it was determined that gas primarily exited the chamber
through the 10 cm exhaust port. Plastic dryer hose was sealed to the chamber exhaust port and
was connected to a 76 cm length of straight PVC pipe. An anemometer was used to measure the
velocity in this 8.3 cm ID pipe. The system ventilation rate (Qg) was calculated using the cross-
sectional area of the pipe (54 cm2) and the measured velocity. The air exchange rate was
determined by dividing the system's ventilation rate by the system volume. The shower system
was well ventilated, with air exchange rates ranging from 12 to 13 air changes per hour (ACH).
The specific air exchange rate for each experiment was measured during the actual experiment,
4.3.3. Parameter Estimation
Each shower experiment was divided into three periods: initial (0 to 1 minute), intermediate
(3.5 to 4.5 minutes), and final (5.75 to 8 minutes). During each period, at least one tracer
reservoir liquid sample, shower outlet sample, and shower gas sample were collected. Chemical
stripping efficiencies and values of KLA were determined for each time period and averaged,
respectively, to obtain final values. Ratios of kg/kl5 kjA, and kgA were estimated based on
averaged values of KLA for each chemical.
4.4. SHOWERRESULTS
Based on the experimental methodology presented in Sections 3.0 and 4.3, the overall
chemical stripping efficiencies and mass transfer coefficients (KLA, ^A, and kgA) for 10 shower
experiments are presented in this chapter. In addition, the effects of liquid temperature, liquid
flowrate, shower spray type, and chemical properties on each response are discussed. The
determination of kg/kj values and associated implications are also presented.
The operating conditions for each experiment are listed in Table 4-1.
4.4.1. Chemical Stripping Efficiencies
Stripping efficiencies for each experimental chemical are presented in Tables 4-2 to 4-6,
respectively. Stripping efficiencies were based on liquid-phase measurements collected from the
tracer reservoir and shower outlet drain. In addition to chemical stripping efficiencies, Tables 4-
2 to 4-6 provide the results of the factorial main effect analysis (see Section 3.7 for
4-5
-------
Table 4-1. Shower experiment operating conditions
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temperature
(°C)
21
22
21
22
35
34
34
36
35
34
Liquid
flowrate
(L/min)
9.1
9.1
6.1
6.1
9.1
9.1
9.1
6.1
6.1
6.1
Gas
flowrate
(L/min)
370
343
360
358
379
354
373
364
371
367
ACH
(1/hr)
13
12
12
12
13
12
13
13
13
13
Spray
type
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
methodology). The three factors of the shower experimental two-level factorial arrays were
shower spray type, liquid flowrate, and liquid temperature. As explained in Section 3.7, the
main effect for a single variable was calculated as the average of the differences between
responses at two levels of the factor of interest. For example, the shower spray effect on
acetone's stripping efficiency may be calculated as:
Corresponding
experiments:
1-2
3-4
5 — Average (6 and 6 replicate) =
7 — Average (8 and 8 replicate) =
Difference in
stripping efficiencies
-2.1%
-0.2 %
1.0%
Average
-0.075%
As shown in Table 4-2, the difference in experimental response was listed twice, once for
each corresponding experiment. Duplicating the listing of each difference in response, however,
does not affect the average value for each variable. The experimental results for Experiments 6
and 6 replicate and Experiments 8 and 8 replicate were averaged, respectively, before applying
factorial analyses. Tables 4-3 to 4-6 follow this same format.
4-6
-------
Table 4-2. Acetone stripping efficiencies for experimental shower
Experiment
#
1
2
3
4
5
6
6 rep.
7
8
8 rep.
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
Stripping
efficiency
(%)
6.3
8.4
9.1
9.3
13
11
12
16
14
15
Average =
Shower
spray
effect3 (%)
-2.1
-2.1
-0.20
-0.20
1.0
1.0
-0.075
Liquid
flowrate
effect"
(%)
-2.8
-0.90
-2.8
-0.90
-3.0
.0
-3.0
.0
-2.4
Liquid
temperature
effect0 (%)
6.7
3.6
6.9
5.7
6.7
3.6
6.9
.7
5.7
a Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
Table 4-3. Ethyl acetate stripping efficiencies for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
Stripping
efficiency
(%)
15
15
20
20
27
28
29
32
33
36
Average =
Shower
spray
effect3 (%)
0
0
0
0
-2.0
-2.0
-3.0
-3.0
-1.3
Liquid
flowrate
effect" (%)
-5.0
-5.0
-5.0
-5.0
-5.0
-6.0
-5.0
-6.0
-5.3
Liquid
temperature
effect0 (%)
12
14
12
15
12
14
12
15
13
" Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
4-7
-------
Table 4-4. Toluene stripping efficiencies for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
Stripping
efficiency
(%)
61
68
63
64
68
75
74
74
73
77
Average =
Shower
spray
effect" (%)
-7.0
-7.0
-1.0
-1.0
-7.0
7 f\
~ 1 .(J
-1.0
1f\
.U
-4.0
Liquid
flowrate
effect" (%)
-2.0
4.0
-2.0
4.0
-6.0
.
-6.0
-
-1.0
Liquid
temperature
effect0 (%)
7.0
7.0
11
11
7.0
11
9.0
° Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
Table 4-5. Ethylbenzene stripping efficiencies for experimental shower
Expteriment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
Stripping
efficiency
'(%)
62
68
63
63
68
75
74
73
72
75
Average =
Shower
spray
effect3 (%)
-6.0
-6.0
0
0
-7.0
-7.0
-1.0
1A
.0
-3.5
Liquid
flowrate
effect" (%)
-1.0
5.0
-1.0
5.0
-5.0
1
-5.0
1
0
Liquid
temperature
effect6 (%)
6.0
7.0 :
10
11
6.0
7
10
11
8.5
* Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
e Liquid temperature effect from cold to warm.
4-8
-------
Table 4-6. Cyclohexane stripping efficiencies for experimental shower
Experiment
#
1
-•- 2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm,
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
Stripping
efficiency
.(%)
65
73
66
66
' 75
77
77
76
75
80
Average =
Shower
spray
effect3 (%)
-8.0
-8.0
0
0
-2.0
-2.0
-2.0
-2.0
-3.0
Liquid
flowrate
effect" C%)
-1.0
7.0
-1.0
7.0
1.0
-1.0
1.0
-1.0
1.0
Liquid
temperature
effect0 (%)
10
4.0
10
12
10
4
10
12
9.0
1 Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
0 Liquid temperature effect from cold to warm.
Stripping efficiencies for acetone ranged from 6.3% to 16%, with the highest value for the
conditions of warm water, low liquid flowrate, and coarse shower spray. The single variable with
the largest effect on acetone's stripping efficiency was liquid temperature, with a main effect of
5.7%. The main effect due to differences in liquid temperature, was calculated by subtracting cold
water stripping efficiencies from corresponding warm water stripping efficiencies. Thus, 5.7%
indicates an absolute increase in stripping efficiency with higher temperature water. The shower
experiments were grouped according to similar liquid temperature, and the following stripping
efficiencies resulted: 8.3% for cold water experiments (Experiments 1 to 4), and 14% for warm
water experiments (Experiments 5 to 8 replicate). This result was expected, owing to the increase
in Henry's law constant with increasing temperature.
For the temperatures listed in Table 4-1, Henry's law constants for acetone ranged from 0.0010
m3,iq/m3gas (21°C, Experiments 1 and 3) to 0.0023 m3liq/m3gas (36°C, Experiment 7).
The second highest main effect involved liquid flowrate with a value of-2.4%. The liquid
flowrate effect was determined by the difference in high flowrate and low flowrate stripping
efficiencies, so a negative effect indicates an increase in stripping efficiency at low flowrates. At
lower shower flowrates, a liquid droplet has a longer residence time in the shower stall, which may
4-9
-------
lead to higher chemical volatilization. The experiments were grouped according to liquid flowrate
and temperature, and the following average stripping efficiencies were calculated: 7.4% for high
flowrate and cold water (Experiments 1 and 2), 9.2% for low flowrate and cold water
(Experiments 3 and 4), 12% for high flowrate and warm water (Experiments 5, 6, and 6 replicate),
and 15% for low flowrate and warm water (Experiments 7, 8, and 8 replicate). Shower spray had a
less significant impact on acetone stripping efficiencies.
Shower Experiments 6 and 8 were replicated. The acetone stripping efficiencies for these two
experiments were compared and the following relative differences calculated: 8.7% for
Experiments 6 and 6 replicate, and 6.9% for Experiments 8 and 8 replicate.
Ethyl acetate stripping efficiencies ranged from 15% to 36% (see Table 4-3). As with
acetone, the highest value corresponded to the conditions of warm water and low flowrate.
However, unlike acetone, the highest stripping efficiency for ethyl acetate occurred during fine
spray conditions. The variable with the highest main effect on ethyl acetate's stripping efficiency
was liquid temperature, with a value of 13%. Ethyl acetate stripping efficiencies were grouped
according to liquid temperature, and the following average values calculated: 18% for cold water
experiments and 31% for warm water experiments. Again, increasing the water temperature
increased ethyl acetate's Henry's law constant, resulting in significantly higher stripping
efficiencies. The Henry's law constant effect is also evident when comparing acetone and ethyl
acetate stripping efficiencies for similar experimental conditions. In all cases, ethyl acetate, which
has a higher Henry's law constant, had higher stripping efficiencies than acetone. For the
temperatures listed in Table 4-1, Henry's law constants (Hc) for ethyl acetate ranged from 0.0041
m3liq/m3gas (21°C, Experiments 1 and 3) to 0.0080 m3liq/m3gas (36°C, Experiment 7), that is,
approximately four tunes that of acetone.
With a main effect value of—53%, liquid flowrate had less than half the impact of water
temperature on ethyl acetate stripping efficiency. When experiments were grouped according to
liquid flowrate and water temperature, the following average values resulted: 15% for cold water
and high flowrate experiments, 20% for cold water and low flowrate experiments, 28% for warm
water and high flowrate experiments, and 34% for warm water and low flowrate experiments.
Again, shower spray had a less significant main effect on ethyl acetate's stripping efficiency.
4-10
-------
For the two replicate experiments, Experiments 6 and 8, the following relative differences
were determined: 3.5% for Experiment 6 and Experiment 6 replicate, and 8.7% for Experiment 8
and Experiment 8 replicate.
As shown in Table 4-4, toluene stripping efficiencies ranged from 61% to 77%. As expected,
the highest toluene stripping efficiencies resulted when warm water was used. The main effect for
liquid temperature was 9.0%. Experiments using cold water had an average stripping efficiency of
64%, and experiments using warm water had an average stripping efficiency of 74%. The gap
between the cold water average stripping efficiency and warm water average stripping efficiency
was much narrower than for acetone and ethyl acetate. For the temperatures listed in Table 4-1,
Henry's law constants for toluene ranged from 0.24 m3liq/m3gas (21°C, Experiments 1 and 3) to 0.38
m3liq/m3gas (3 6°C, Experiment 7).
The second largest main effect for toluene stripping efficiencies, unlike those for acetone and
ethyl acetate, was the type of shower spray, with a value of -4.0%. Interestingly, the magnitude of
the shower spray, main effect was highly dependent on liquid flowrate. The difference in stripping
efficiency between shower spray types at high flowrates was —7.0%, but at low flowrates the
difference was only -1.0%. Interaction between these two variables is likely to influence the
magnitude of a chemical's liquid-phase mass transfer coefficient (kj). Thus, the associated effects
of liquid flowrate and shower spray will have the greatest effect on chemicals dominated by
liquid-phase resistance to mass transfer (toluene, ethylbenzene, and cyclohexane). Toluene
stripping efficiencies were grouped according to the two largest main effects, water temperature
and shower spray type, and the following average values were calculated: 62% for cold water and
coarse spray (Experiments 1 and 3), 66% for cold water and fine spray (Experiments 2 and 4),
71% for warm water and coarse spray (Experiments 5 and 7), and 75% for warm water and fine
spray (Experiments 6, 6 replicate, 8, and 8 replicate).
Replicate experimental results led to a 1.3% relative difference hi toluene stripping
efficiencies for Experiments 6 and 6 replicate, and 5.3% relative difference for Experiments 8 and
8 replicate.
4-11
-------
Ethylbenzene stripping efficiencies ranged from 62% to 75% (see Table 4-5). This range was
similar in magnitude to the range of stripping efficiencies reported for toluene. As discussed in
Section 3.2.1, toluene and ethylbenzene have similar Henry's law constants (for the temperatures
listed in Table 4-1, ethylbenzene has Henry's law constants between 0.26 m3Iiq/m3gas and 0.57
nr'ijq/m3^), and thus should yield similar volatilization results. On an experiment-by-experiment
basis, toluene and ethylbenzene stripping efficiencies were nearly identical. The largest relative
deviation in stripping efficiencies for the two compounds was less than 3% (Experiment 8
replicate). It should also be noted that the stripping efficiencies for toluene and ethylbenzene were
significantly higher than those observed for acetone and ethyl acetate. Again, an increase in
Henry's law constant led to an increase in chemical stripping efficiencies.
As expected, ethylbenzene had main effects similar to those of toluene. Grouping stripping
efficiencies based on water temperature yielded the following averages: 64% for cold water
experiments and 73% for warm water experiments. Separating the liquid temperature groups to
account for shower spray type resulted in the following average values: 63% for cold water and
coarse spray, 66% for cold water and fine spray, 71% for warm water and coarse spray, and 74%
for warm water and fine spray.
Relative differences in stripping efficiency for replicate experiments were 1.3% for
Experiments 6 and 6 replicate, and 4.1% for Experiments 8 and 8 replicate.
Finally, cyclohexane stripping efficiencies ranged from 65% to 80% (see Table 4-6). For
similar experimental conditions, cyclohexane consistently had the highest stripping efficiency of
the five experimental tracers. The largest main effect was liquid temperature with a value of 9.0%.
Following the format for previous tracers, the average cold water stripping efficiency was 68%,
and the average warm water stripping efficiency was 77%. Similar to toluene and ethylbenzene,
shower spray type had the second highest main effect with a value of-3.0%. Experimental results
were regrouped according to shower spray type and water temperature, and the following averages
were calculated: 66% for cold water and coarse spray, 70% for cold water and fine spray, 76% for
warm water and coarse spray, and 77% for warm water and fine spray. This second regrouping
did not yield results significantly different from the first set of averages for cold and warm water,
4-12
-------
and was thereby unnecessary. For the temperatures listed in Table 4-1, Henry's law constants for
cyclohexane ranged from 6.3 m3liq/m3gas (21°C, Experiments 1 and 3) to 10 m3liq/m3gas (36°C,
Experiment 7).
Unlike the other chemical tracers, the liquid flowrate main effect on cyclohexane's stripping
efficiencies was positive, indicating a decrease in stripping efficiency with decreasing flowrate. A
specific reason for this trend could not be identified.
Replicate experimental stripping efficiencies had a relative difference of 0% for Experiments
6 and 6 replicate and 6.5% for Experiments 8 and 8 replicate.
j
An attempt was made to compare the chemical stripping efficiencies described above with
those reported by other researchers who used similar operating conditions and chemical tracers. A
summary of previous research related to volatilization in showers was presented in a Phase I report
to EPA as part of this project (Corsi et al, 1996) and are also given in the database in the
Appendix. Additional papers have been reviewed since the Phase I report was submitted (e.g.,
Giardino and Andelman [ 1996]), and all of these have been added to the database.
Previous researchers have not studied chemicals with Henry's law constants as low as
acetone. Thus, the results described herein are unique for this compound and extend the range of
chemical volatilities to values much lower than those previously reported.
Overall, Giardino and Andelman (1996) used operating conditions most similar to those in
this study and will serve as the primary basis of comparison. Giardino and Andelman studied
emissions of trichloroethene (TCE), chloroform (CHC13), and l,2-dibromo-3-chloropropane
(DBCP) in a 1.5 m3 experimental shower. As in the results of this study, they determined that
water temperature had a dominant effect on the total release of each tracer chemical.
Giardino.and Andelman's Experiment 17 included an air exchange rate of 12.3/hour, water
flowrate of 5 L/minute, and water temperature of 30°C. For these conditions, the stripping
efficiency of DBCP, which has the lowest Henry's law constant of any chemical tested to date for
showers, was only 17%. For this study, Experiment 8 replicate included operating conditions
4-13
-------
similar to those reported above (air exchange rate = 13/hour; water flowrate = 6.1 L/minute; water
temperature = 34°C). The corresponding stripping efficiency for ethyl acetate, a chemical with a
Henry's law constant at 34°C (slightly lower than that of DBCP at 30°C), was over twice (36%)
the value reported by Giardino and Andelman for DBCP. Ethyl acetate's Henry's law constant is
similar to that of DBCP, and thus differences in stripping efficiency between DBCP and ethyl
acetate cannot be accounted for entirely by water temperature. Differences are likely due to
differences in commercial showerheads that were used, as well as subsequent differences in
droplet sizes and velocities.
Giardino and Andelman (1996) also studied TCE, which has a Henry's law constant
approximately 25% greater than that for ethylbenzene, at 22°C. Thus, TCE would be expected to
have slightly greater stripping efficiencies for similar operating conditions. Giardino and
Andelman reported a TCE stripping efficiency of 60% for their Experiment 2 (air exchange rate =
10.8/hour; water flowrate = 5.1 L/minute; water temperature = 22°C). In this study, the stripping
efficiency for ethylbenzene was slightly higher (63%) for similar conditions (Experiment 3; air
exchange rate = 12/hour; water flowrate = 6.1 L/minute; water temperature = 21°C). For a second
experiment involving a higher water flowrate (10 L/minute), Giardino and Andelman observed a
TCE stripping efficiency of 57%. For similar experimental conditions (Experiments 1 and 2 of
this study), the stripping efficiency for ethylbenzene was observed to be 62% (coarse spray) and
68% (fine spray). In an earlier study, Giardino et al. (1992) observed TCE stripping efficiencies of
59% to 67% for similar operating conditions.
McKone and Knezovich (1991) also studied stripping efficiencies for TCE in an experimental
shower. One of their operating conditions (air exchange rate = 12/hour; liquid flowrate —9.5
L/minute; water temperature = 22°C) was nearly identical to those used in Experiments 1 and 2 of
this study. The stripping efficiency for TCE was reported to be 58%, consistent with Giardino and
Andelman (1996) and Giardino et al. (1992), and slightly lower than those obtained for
ethylbenzene hi this study. The differences in stripping efficiencies between TCE and
ethylbenzene could easily be caused by differences in hydrodynamic conditions associated with
water flowrate and shower configurations, as well as experimental errors associated with each
study.
4-14
-------
Finally, several researchers used chemicals with relatively high Henry's law constants-(> 2.0
m3liq/m3gas) in shower experiments. It was expected that these higher volatility chemicals would
have similar stripping efficiencies because of the associated insignificance of gas-phase resistance
to mass transfer. For example, Bernhardt and Hess (1995) studied stripping efficiencies for radon
in household showers. Radon has a slightly lower Henry's law constant than cyclohexane, but
both compounds should be dominated by liquid-phase resistance to mass transfer. For a water
temperature of 23°C and liquid flowrate of 5.7 L/minute (gas exchange rate in the shower stall was
not measured), the stripping efficiency for radon was reported to be 78%. For similar operating
conditions (Experiments 3 and 4 of this study), the stripping efficiency for cyclohexane was
determined to be 66%. The range of radon stripping efficiencies reported by Bernhardt and Hess
was 57% to 88%. Cyclohexane stripping efficiencies for this study ranged from 65% to 80%.
At liquid flowrates of 2 to 4 L/minute, Giardino and Hageman (1996) measured radon
stripping efficiencies ranging from 67% to 70%. Studies with unknown operating parameters led
to observed radon stripping efficiencies of 63% to 71% (Gesell and Prichard, 1980; Hess et al.,
1982; Hopke et al., 1995; Partridge, 1979).
Tancrede et al. (1992) measured the stripping efficiencies of five experimental chemicals
including carbon tetrachloride (CC14), which has a Henry's law constant of 2.3 m3liq/m3gas at 42°C.
The chemical stripping efficiency for CC14 was 59% at a liquid flowrate of 9.7 L/minute and 77%
for a liquid flowrate of 13 L/minute. Again, these results are consistent with those observed for
other chemicals with relatively high Henry's law constants.
It is clear from this study, as well as several others reported in the literature, that for the same
operating conditions stripping efficiency increases with increasing Henry's law constant. It is also
evident that chemicals of sufficiently high Henry's law constant have comparable stripping
efficiencies for similar operating conditions. Because the conditions used in this study should
represent a reasonable spectrum of those associated with residential showering, an average
stripping efficiency was determined for each chemical tracer and is plotted in Figure 4-3 as a
function of Henry's law constant at 25°C. This plot may be used as a screening tool for
4-15
-------
approximating chemical stripping efficiencies, given knowledge of that chemical's Henry's law
constant at 25°C, the temperature for which Henry's law constants are most widely reported. The
best-fit line associated with the averaged data in Figure 4-3 stems from the following relationship:
= 7.5 •
68.2
(4-1)
where
HL
= Henry's law constant for chemical of interest (L3Iiq/L3gas).
Although Equation 4-1 provides a relationship for chemical stripping efficiencies averaged
over a wide range of shower operating conditions, it does provide insight into differences in
potential stripping efficiencies for various types of compounds. However, application of Equation
4-1 to chemicals with Henry's law constants beyond the range of those used to develop this
relationship is not recommended.
Equation 4-1 can be rearranged to solve for the value of Henry's law constant that leads to
specific stripping efficiencies. For example, the value of Hc that leads to r\ = 55% is 0.19
m3Hq/ni3gas. This Henry's law constant is consistent with reported values for chloroform at 25°C
Henry's LawConstant (m Hq/m gas)
Figure 4-3. Relationship between Henry's law constant and average stripping efficiency.
4-16
-------
(Howard, 1990), a common disinfection by-product. Tancrede et al. (1992) reported chloroform
stripping efficiencies ranging from 52% to 53%, Giardino and Andelman (1991) reported a value
of 55%, and Giardino and Andelman (1996) reported chloroform stripping efficiencies ranging
from 44% to 52%, all in good agreement with Equation 4-1.
4.4.2. KLA Values
Values of KLA for each chemical tracer are reported in Tables 4-7 to 4-11. The determination
of values of KLA was based on liquid-phase data for all chemicals. Tables 4-7 through 4-11 have a
format similar to that of Tables 4-2 to 4-6, except the main effects are based on values of KLA.
Values of KLA for acetone ranged from 1.4 to 3.7 L/minute (see Table 4-7). The highest
value corresponded to the experimental conditions of warm water, high flowrate, and fine shower
spray. The largest main effect was liquid flowrate, with a value of 0.93 L/minute. In a manner
similar to that for stripping efficiency results, KLA values can be grouped according to liquid
flowrate, resulting in the following average values: 2.9 L/minute for high flowrate and 2.0
L/minute for low flowrate.
Liquid temperature had the second highest main effect on KLA values for acetone. The liquid
temperature main effect was 0.83 L/minute, which indicated an increase in KLA with increased
temperature. As expected from its greater surface to volume ratio, fine shower spray was
determined to increase stripping efficiencies more than did coarse spray.
Values of KLA for the replicate experiments were also compared. For Experiments 6 and 6
replicate, the relative difference in values of KLA was 8.5%. For Experiments 8 and 8 replicate,
the relative difference in values of KLA was 8.3%.
Measured and predicted liquid-phase and gas-phase concentrations of acetone for
Experiment 7 are presented in Figure 4-4, and are representative of other experiments. The
operating conditions used in Experiment 7 were warm water, low flowrate, and coarse shower
spray. As described in Section 4.3.3, each shower experiment was divided into three separate
4-17
-------
Table 4-7. Acetone KLA values for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
KLA
(L/min)
1.8
3.0
1.4
1.5
2.8
3.4
3.7
2.2
2.3
2.5
Average =
Shower
spray
effect3
(L/min)
-1.2
-1.2
-0.10
-0.10
-0.80
Oon
.oil
-0.20
-0.20
-0.58
Liquid
flowrate
effect"
(L/min)
0.40
1.5
0.40
1.5
0.60
1.2
0.60
1.2
0.93
Liquid
temperature
effect0
(L/min)
1.0
0.60
0.80
0.90
1.0
0.6
0.80
0.9
0.83
* Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
Table 4-8. Ethyl acetate KLA values for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
KLA
(L/min)
2.9
4.0
2.3
2.5
5.5
6.9
6.7
3.8
4.7
5.3
Average =
Shower
spray
effect3
(L/min)
-1.1
-1.1
-0.20
-0.20
-1.3
1 "5
i.j
-1.2
1O
.2
-0.95
Liquid
flowrate
effectb
(L/min)
0.60
1.5
0.60
1.5
1.7
1.8
1.7
1.8
1.4
Liquid
temperature
effect0
.(L/min)
2.6
2.8
1.5
2.5
2.6
2.8
1.5
2.5
2.4
* Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
0 Liquid temperature effect from cold to warm.
4-18
-------
Table 4-9. Toluene KLA values for experimental shower
Experiment
'#-
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
KLA
(L/min)
8.8
11
6.2
6.4
11
13
12
8.4
8.1
9.2
Average =
Shower
spray
effect3
(L/min)
-2.2
-2.2
-0.20
-0.20
-2.0
-2.0
-0.30
-0.30
-1.2
Liquid
flowrate
effect15
(L/min)
2.6
4.6
2.6
4.6
2.6
4.3
2.6
4.3
3.5
Liquid
temperature
effect0
(L/min)
2.2
2.0
2.2
2.2
2.2
2
2.2
2.2
2.2
a Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
Table 4-10. Ethylbenzene KLA values for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
KLA
(L/min)
8.9
11
6.0
6.2
11
13
12
8.2
7.9
8.8
Average =
Shower
spray
effect3
(L/min)
-2.1
-2.1
-0.20
-0.20
-2.0
.0
-0.20
.20
-1.1
Liquid
flowrate
effect"
(L/min)
2.9
4:8
2.9
4.8
2.8
4.6
2.8
4.6
3.8
Liquid
temperature
effect6
(L/min)
1.1
2.0
2.2
2.2
1.1
2
2.2
2.2
2.1
a Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
4-19
-------
Table 4-11. Cyclohexane KLA values for experimental shower
Experiment
#
1
2
3
4
5
6
6 replicate
7
8
8 replicate
Liquid
temp.
Cold
Cold
Cold
Cold
Warm
Warm
Warm
Warm
Warm
Warm
Liquid
flowrate
High
High
Low
Low
High
High
High
Low
Low
Low
Shower
spray
Coarse
Fine
Coarse
Fine
Coarse
Fine
Fine
Coarse
Fine
Fine
KLA
(L/min)
9.6
12
6.5
6.7
13
14
13
8.6
8.4
9.9
Average =
Shower
spray
effect3
(L/min)
-2.4
• -2.4
-0.20
-0.20
-1.0
1A
.U
-0.60
0/~r\
.OU
-1.1
Liquid
flowrate
effect"
(L/min)
3.1
5.3
3.1
5.3
4.4
4.8
4.4
4.8
4.4
Liquid
temperature
effect0
(L/min)
3.4
2.0
2.1
2.5
3.4
2.1
2.0
2.5
2.5
* Shower spray effect from fine to coarse.
b Liquid flowrate effect from low to high.
c Liquid temperature effect from cold to warm.
- 0.08
H i 1 H 1—: 1 h
60 105 150 195 240 285 330 375 420 465
Time (seconds)
0.00
Meas ured Res ervoir Liquid Values
Model Predicted Liquid Values
.ModelPredicted Gas Values
Measured Liquid Values
Measured Gas Values
Figure 4-4. Acetone experimental data for Experiment 7.
4-20
-------
periods: initial, intermediate, and final. As shown in Figure 4-4, each experimental period
consisted of a liquid sample collected from the tracer reservoir, an outlet liquid sample, and a gas
sample. For each period, the shower outlet concentration in both the liquid and gas phases may be
estimated using the shower mass balance models (Equations 2-28 and 2-30). To determine the
best value of KLA for the shower model, the residuals between the measured and predicted
concentrations were minimized using the method described in Section 3.6.2. Two liquid samples
were collected in the initial period for one gas sample. Thus, for this period the liquid-phase
residual was based on the average of two measured liquid samples and a model-predicted value.
For Experiment 7 shown in Figure 4-4, the best-fit value of KLA for acetone was 2.2 L/minute.
During each experiment, the chemical concentration in the tracer reservoir was relatively '
constant between each period. For acetone, the liquid-phase concentration measured in the shower
drain tended to increase with experimental time, as mass accumulated in the shower atmosphere
(gas phase). This accumulation resulted in a decreased chemical concentration driving force. The
acetone gas-phase concentration continually increased during each experiment.
Values of KLA for ethyl acetate ranged from 2.3 to 6.9 L/minute, approximately 1.6 times
greater than values reported for acetone. The highest value was for the experimental conditions of
warm water, high flowrate, and fine spray. The largest main effect was liquid temperature, with a
value of 2.4 L/minute. The average cold water value of KLA for ethyl acetate was 2.9 L/minute,
and the average warm water value of KLA was 5.5 L/minute. Again, values of KLA tended to
increase with increasing flowrate and fine spray.
Replicate values of KLA for ethyl acetate had a relative difference of 2.9% for Experiments 6
and 6 replicate, and 12% for Experiments 8 and 8 replicate.
Experimental results for ethyl acetate during shower Experiment 7 are presented in Figure 4-
5. The value of KLA of 3.8 L/minute for this experiment was determined by minimizing the
residuals between the measured liquid concentration data points and predicted liquid
concentrations. As shown in Figure 4-5, for relatively constant inlet liquid concentrations
(measured tracer reservoir liquid values), the measured outlet liquid-phase concentrations
4-21
-------
increased with time. This increase in concentration reflected the decreasing chemical driving force
as mass accumulated in the shower stall. - As with acetone, ethyl acetate gas-phase concentrations
increased with experimental time, rapidly within the first 150 seconds and more gradually
thereafter. All gas-phase data in experimental plots represent measurements taken at gas sample
port#l.
Values of KLA for toluene ranged from 6.2 to 13 L/minute (see Table 4-9). Similar to acetone
and ethyl acetate, the operating conditions of warm water, high liquid flowrate, and fine shower
spray resulted in the highest value. However, for toluene the highest main effect was not for water
temperature, but rather liquid flowrate. This trend is consistent with a shift from gas-phase
resistance dominating volatilization of acetone and ethyl acetate to liquid-phase resistance
dominating for toluene, ethylbenzene, and cyclohexane. Because water temperature has its
greatest influence on Henry's law constant, for higher values of Hc the effect of temperature is
significantly reduced as the Cg/Hc term on the right-hand side of Equation 2.27 is reduced.
Consequently, hydrodynamic effects on k] and A become more important.
A main effect value of 3.5 L/minute indicated that toluene KLA values increased with
0.10
0.00
60 105 150 195 240 285 330 375 420 465
Time (seconds)
Measured Res ervoir Liquid Values
Model Predicted Liquid Values
. Model Predicted Gas Values
Measured Liquid Values
Measured Gas Values
Figure 4-5. Ethyl acetate experimental data for Experiment 7.
4-22
-------
increasing liquid flowrate. Values of KLA for toluene were grouped according to high and low
flowrate, and the following average values were calculated: 11 L/minute for high flowrate
experiments (Experiments 1, 2, 5, 6, and 6 replicate) and 7.7 L/minute for low flowrate
experiments (Experiments 3, 4, 7, 8, and 8 replicate).
The second largest main effect was for liquid temperature, with a value of 2.2 L/minute.
When the experiments were regrouped using liquid flowrate and liquid temperature, the following
averages resulted: 9.9 L/minute for cold water and high flowrate (Experiments 1 and 2), 6.3
L/minute for cold water and low flowrate (Experiments 3 and 4), 12 L/minute for warm water and
high flowrate (Experiments 5, 6, and 6 replicate), and 8.6 L/minute for warm water and low
flowrate (Experiments 7, 8 , and 8 replicate). Fine shower spray resulted in higher values of KLA
for toluene than coarse spray as a result of the increased total surface area for the liquid phase.
Toluene results for Experiment 7 are presented in Figure 4-6. Differences between toluene
concentrations in the tracer reservoir concentrations and shower outlet were significantly greater
than differences shown in Figures 4-4 and 4-5 for acetone and ethyl acetate, respectively. This
larger difference reflects the greater chemical volatilization rate for toluene, which is less affected
Time (seconds)
Meas ured Res ervoir Liquid Values
Model Predicted Liquid Values
.ModelPredicted Gas Values
Measured Liquid Values
Measured Gas Values
Figure 4-6. Toluene experimental data for Experiment 7.
4-23
-------
by an approach to chemical equilibrium, that is, reduction in the concentration driving force
between water and air, and gas-phase resistance to mass transfer. For toluene, the ratio of gas
concentration to Henry's law constant was always small relative to inlet or outlet water
concentration (Cg/Hc« C, hi Equation 2.27). This condition led to nearly constant values of
toluene concentration in water draining from the shower and constant stripping efficiencies during
the course of an experiment, trends that were also observed for ethylbenzene and cyclohexane.
Values of KLA for ethylbenzene ranged from 6.0 to 13 L/minute (see Table 4-10). As
expected, this range is similar in magnitude to that of toluene. Ethylbenzene also shared main
effects similar to those calculated for toluene. Grouping ethylbenzene KLA values by high and
low flowrate resulted in the following average values: 11 L/minute and 7.4 L/minute,
respectively.
Values of KLA for ethylbenzene may also be grouped according to liquid flowrate and liquid
temperature. Average values were 10 L/minute for high flowrate and cold water, 6.1 L/minute for
low flowrate and cold water, 12 L/minute for high flowrate and warm water, and 8.3 L/minute for
low flowrate and warm water.
0.06
o.oo
150 195 240 285 330 375 420 465
Time (seconds)
Measured Reservoir Liquid Values
Model Predicted Liquid Values
.ModelPredicted Gas Values
Measured Liquid Values
Measured Gas Values
Figure 4-7. Ethylbenzene experimental data for Experiment 7.
4-24
-------
Ethylbenzene data for Experiment 7 are plotted in Figure 4-7. Chemical concentration
values and trends follow those discussed for toluene. Both chemicals had a value of KLA of 13
L/minute for this experiment.
Finally, values of KLA for cyclohexane ranged from 6.5 to 14 L/minute (see Table 4-11).
The fact that cyclohexane has a significantly higher Henry's law constant than either toluene or
ethylbenzene but its values of KLA were only slightly higher suggests that gas-phase resistance to
mass transfer was small for each of these three tracers. Following the trend of toluene and
ethylbenzene, cyclohexane also had the highest main effect value associated with liquid flowrate,
with a value of 4.4 L/minute. Average values of KLA based on liquid flowrate were 12 L/minute
for high flowrate and 8.0 L/minute for low flowrate.
Cyclohexane data are plotted in Figure 4-8 for Experiment 7. Again, for relatively constant
inlet liquid concentrations, the outlet liquid-phase concentrations were consistent with one another.
Cyclohexane gas-phase concentrations increased at a consistent rate throughout each experiment.
0.025
60 105 150
195 240 285
Time (seconds)
330 375 420 465
Measured Reservoir Liquid Values
Model Predicted Liquid Values
Model Predicted Gas Values
X
Measured Liquid Values
Measured Gas Values
Figure 4-8. Cyclohexane experimental data for Experiment 7.
4-25
-------
To calculate KLA for each chemical tracer using the shower mass balance models (Equations
2-28 and 2-30), the gas phase was assumed to simulate a well-mixed reactor. To check the
validity of this assumption, gas-phase samples were collected at three locations within the shower
atmosphere as shown in Figure 4-1. Based on the percent difference between measured gas-phase
concentrations at each sample port, the shower stall appeared to be relatively well mixed. The
average of percent differences (absolute values) between gas-phase samples for acetone were 18%
when comparing sample port #1 and sample port #2,16% when comparing sample port #2 and
sample port #3, and 16% when comparing sample port #1 and sample port #3. The concentration
differences between sample ports appeared to be random between experiments; that is, the relative
differences were both positive and negative. In addition, 85% of compared samples were within
0.02 mg/L. The average percent differences for ethyl acetate were 20% when comparing sample
ports #1 and #2,17% when comparing sample ports #2 and #3, and 20% when comparing sample
ports #1 and #3. The average percent differences for the remaining compounds ranged from 18%
to 30%. Over 93% of the compared gas-phase samples for toluene and ethylbenzene were within
0.02 mg/L, and over 88% of the compared gas-phase samples for cyclohexane were within 0.002
mg/L. Small deviations from this well mixed assumption should have no effect on experimentally
determined values of KLA for toluene, ethylbenzene, and cyclohexane, in that Cg/H0« C, for
these chemicals.
4.4.3. Liquid-and Gas-Phase Mass Transfer Coefficients
For future model applications, it is valuable to separate KLA into liquid- and gas-phase
components, that is, k,A and kgA, and to predict kg/T^ values for different operating conditions. For
a specific system, values of kg/k, should not vary significantly between volatile chemicals (Munz
and Roberts, 1989). Values of k,A and kgA for each chemical tracer are listed in Table 4-12. A
single value of kg/k] is presented based on all chemical tracer experimental KLA values and
physicochemical properties, as described in Section 3.6.3. The relative difference between
replicate experiments was 15% for Experiments 6 and 6 replicate and 3.6% for Experiments 8 and
8 replicate.
With use of the factorial analysis described in Sections 3.7 and 4.4.1, the impact of shower
operating conditions on k,A and kgA was investigated. As with KLA, the most significant
4-26
-------
operating condition affecting k,A was liquid fiowrate, except for ethyl acetate, which was most
affected by temperature. The most significant factor affecting kgA was liquid fiowrate, this time
for all chemicals. As expected, temperature generally had a greater relative impact on k,A than
kgA.
As shown in Table 4-12, the ratio of kg/k, for showers ranged from 110 to 223, with an
average value of 156. This value is consistent with the typical value of 150 reported by Mackay et
al. (1979). However, Little (1992) reported three values of kg/k, for showers based on other
researchers' work (Giardino and Andelman, 1991; Tancrede et al., 1992). These values were 13
for a liquid temperature of approximately 44°C and liquid fiowrate of 5 L/minute, 22 for a liquid
temperature of 42°C and a liquid fiowrate of 13 L/minute, and 17 for a liquid temperature of 33°C
and liquid fiowrate of 14 L/minute.
An important parameter that influences the back-calculation of kg/k, is the Henry's law
constant for each chemical. As discussed in Section 4.2.1, there is uncertainty associated with
Henry's law constants for chemicals, especially at elevated temperatures. Increasing the Henry's
law constant for toluene in Experiment 7 by 40% results in a 1.1 % decrease in KLA. Thus, values
of KLA for chemicals of higher volatility are less sensitive to changes in Henry's law constant.
However, this is not the case for chemicals such as acetone or ethyl acetate.
Table 4-12. Liquid- and gas-phase mass transfer coefficients for shower experiments
Experiment
#
1
2
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
k,A
(L/min)
13
7.3
9.0
9.1
9.6
16
8.1
11
11
12
kgA
(L/min)
1,986
1,111
1,380
1,395
1,468
3,519
1,807
2,434
2,384
2,652
kp/ki
153
223
4-27
-------
Table 4-12. Liquid- and gas-phase mass transfer coefficients for shower experiments (continued).
Experiment
#
3
4
5
6
6 replicate
7
8
8 replicate
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
k,A
(L/min)
8.6
5.1
6.4
6.2
6.5
8.8
5.3
6.5
6.3
6.7
14
12
11
11
13
16
14
14
13
14
15
13
13
12
13
11
8.2
8.6
8.3
8.6
9.6
9.0
8.3
8.0
8.4
11
10
9.3
8.9
9.9
kgA
(L/min)
1,723
1,030
1,274
1,234
1,305
1,720
1,031
1,275
1,232
1,309
1,548
1,322
1,223
1,188
1,439
2,095
1,852
1,776
1,708
1,786
2,316
1,945
1,930
1,855
1,950
1,169
901
949
917.
943
1,380
1,292
1,189
1,139
1,203
1,507
1,443
1,291
1,227
1,366
yk,
900
A* \j\j
195
111
131
153
110
143
138
4-28
-------
Acetone
Ethyl Acetate
Toluene
Chemicals
Ethylbenzene Cyclohexane
I Liquid-Phase Resistance (min/L) H Gas-Phase Resistance (min/L)
Figure 4-9. Resistances to mass transfer for each chemical in Experiment 7.
Increasing the Henry's law constants of these two chemicals by 40% results in a 17% decrease in
KLA for ethyl acetate and 23% decrease in KLA for acetone. The decrease in KLA for these two
compounds then results in a best-fit kg/k, value of 46, 58% of the value reported in Table 4-12 for
Experiment 7. Interestingly, the best-fit kg/k, value using only toluene, ethylbenzene, and
cyclohexane data from Experiment 7 was 116.
, Liquid and gas-phase mass transfer coefficients may also be used to determine the relative
importance of liquid and gas-phase resistances to mass transfer for specific chemicals and
operating conditions. As shown in Equation 2.5, the overall resistance to mass transfer (1/KLA)
may be written as the sum of liquid-phase resistance to mass transfer (l/k,A) and gas-phase
resistance to mass transfer (l/kgA»Hc). These resistances are shown graphically in Figure 4-9 for
each chemical in Experiment 7. As shown in Figure 4-9, the overall resistance to mass transfer for
acetone is dominated by resistance in the gas phase. The overall resistance to mass transfer for
ethyl acetate is distributed relatively equally between liquid-phase resistance and gas-phase
resistance. Finally, the gas-phase resistances to mass transfer for toluene, ethylbenzene, and
cyclohexane are insignificant relative to their respective liquid-phase resistances to mass transfer.
4-29
-------
4.4.4. Mass Closure
For shower experiments, mass closure values as defined by Equation 3.12 ranged from 96%
to 103% for acetone, 98% to 108% for ethyl acetate, 71% to 90% for toluene, 54% to 73% for
ethylbenzene, and 40% to 74% for cyclohexane. The more volatile chemicals (toluene,
ethylbenzene, and cyclohexane) tended to achieve mass closure values less than 100%. This may
have been due to the dissolution problems described in Section 3.4.2. A separate calibration curve
was developed to assess this effect, based on a 4-day standard calibration period, that is, allowing
chemicals to dissolve hi the Tedlar™ bag for 4 days instead of 1 day. The resulting mass closures ,
improved for toluene (77% to 106%), ethylbenzene (64% to 92%), and cyclohexane (66% to
85%).
Previous researchers (Keating and McKone, 1993; Keating et al, 1997; Tancrede et al., 1992)
have also observed differences hi predicted gas-phase concentrations and measured gas-phase
concentrations for volatile chemicals. It has often been suggested that there exists a second
compartment in the shower system that acts as a chemical sink. Keating and McKone discussed
the possibility that the second-compartment effect could be accounted for by one to all of the
following: incomplete mixing within the shower stall, sorption of chemicals onto surfaces, and/or
scavenging of chemicals by aerosols. A number of tests were completed to investigate these
possibilities. Cyclohexane is used as the example chemical, because it had the most problems
meeting the mass closure requirements.
First, a shower experiment with clean (no chemicals) warm water was completed. At the end
of the experiment, sponges were used to soak up the water collected in known areas on the
different types of surfaces within the shower stall (plastic-coated fiberglass wall and floor,
stainless steel ceiling, and Tedlar™ shower curtain). The sponges were weighed before and after
water collection to estimate total volume of water collected on each surface type. Based on this
experiment, the total water volume present on surfaces at the end of an experiment was
approximately 0.2 L. Using the gas-phase concentration measured for each chemical and
assuming that equilibrium conditions hold at the wetted surface, the expected chemical
concentration of the wall surface water may be calculated. For example, the maximum
concentration measured for cyclohexane was approximately 0.01 mg/L. For a Henry's law
4-30
-------
constant of 10 m3liq/m3gas, the expected liquid-phase concentration would be 0.001 mg/L. For a
total wetted surface volume of 0.2 L, the total mass to be added to the mass closure assessment
would be 0.0002 mg. The total mass of cyclohexane in the shower stall gas phase was 17 mg.
Therefore, the wetted surfaces were not likely to cause the difference between predicted and
measured gas-phase concentrations.
As discussed in Section 4.3.1, a gas-phase sample was collected at the end of an experiment
with no water flowing through the system. This sample was collected to determine the extent of
chemical desorption from shower stall surfaces resulting from chemical adsorption during an
experiment. Measured chemical concentrations were consistently lower than predicted values
based on decay due to ventilation.
As explained in Section 4.4.2, gas-phase samples were collected at different locations within
the shower stall. In general, the shower stall was determined to be well mixed. For mass closure
calculations, concentrations measured at the system's exhaust port were used, and for the most part
appeared to be representative of gas-phase concentrations within the shower stall.
Liquid droplet sizes produced by the experimental showerhead were not measured, making it
difficult to predict the aerosol scavenging effect. On the basis of other shower studies (Keating
and McKone, 1993), it is expected that this phenomenon did not contribute significantly to the
chemical "sink" effect.
When possible, mass closures were determined for previously reported studies. Results were
reported in the Phase I report (Corsi et al., 1996) of this project and in the Appendix to this report.
In general, the mass closures determined for this study compared favorably with previously
reported shower experiments and in most cases improved upon mass closures for chemicals with
similar Henry's law constants.
4-31
-------
-------
5. DISHWASHER EXPERIMENTS
In dishwasher operation, chemicals originating in tap water can enter the machine and
then be emitted to indoor air during one to five cycles (depending on wash cycle option
selected). These cycles have similar operating characteristics. Thus, only a single cycle was
studied experimentally. For this experimental cycle, a wide range of operating conditions was
applied to adequately characterize the features associated with all of the cycles.
5.1. EXPERIMENTAL SYSTEM
A Kenmore™ dishwasher (Model No. 17651) was purchased and used for all
experiments. The experimental system is illustrated in Figure 5-1. The experimental dishwasher
had an interior volume of 188 L. It had five different types of wash/rinse cycle: Quick Rinse,
China Light, Water Miser, Normal, and Pots and Pans. These cycle options differed only by the
number of fills, that is, total volume of water used and length of operation.
The sequence of operation was as follows: the dishwasher was directly plumbed to a
pressurized hot water line. At the start of each cycle, water entered the dishwasher from the hot
water line at a flowrate of 4.1 L/minute. Twenty seconds after starting the fill, water from the
basin pool was pumped to the rotary arm, which began spinning and spraying water throughout
the dishwasher headspace. Water continued to enter the dishwasher from the hot water line for
99 seconds, such that a total of 6.8 L of water was present in the dishwasher. During the wash
cycle, detergent was released from the holder as the dishwasher filled. At the end of each cycle,
the water was pumped from the basin to a drain. Once all wash and rinse cycles were completed,
there was an approximate 3 0-minute drying time.
The dishwasher was configured to allow for the required variable measurements to solve
the dishwasher mass balance equations (Equations 2-23 and 2-24). A liquid sample port was
installed at the bottom of the dishwasher. A 15 cm length of 0.635 cm OD Teflon™ tubing and a
Teflon™ sample valve were connected to this port. The port inlet was observed to be submerged
at all sampling times, and the residence time of the sample tube was estimated to be
approximately 2 seconds. Liquid samples were collected as described in Section 3.3.1.
5-1
-------
Vent
Gas Sample
Port
Lower Spray Arm
Float
Liquid Temperature
Figure 5-1. Dishwasher experimental system.
Gas samples were collected on sorbent tubes as described in Section 3.3.2. The sorbent
tube was attached to a sample port located in the headspace of the dishwasher. A 2.5 cm OD
Teflon™ tube was connected to the port on the inside of the dishwasher. Sample flowrates were
in the range of 0.2 to 0.4 L/minute, as measured using a bubble flowmeter, and sampling times
were approximately 30 seconds.
In addition to the liquid and gas sample ports, the dishwasher was configured to allow for
liquid temperature monitoring. A thermocouple probe was submerged in the dishwasher pool
and connected to a digital monitor to allow for constant temperature readings.
5.2. EXPERIMENTAL DESIGN
Dishwasher operating variables included water temperature, dish-loading pattern, and use
of detergent (wash versus rinse portions). Experiments were designed to study the effects of
these parameters on chemical volatilization rates using the 2 x 2 x 2 factorial array shown in
Figure 5-2. The numbers hi Figure 5-2 correspond to the experiment number that was
completed with the associated operating conditions. For example, Experiment 1 had the
operating conditions of a rinse cycle (no detergent present), empty machine (no dishes present),
5-2
-------
Rinse —
Wash
Empty
- Water Temperature ~ 54 C
4} - Water Temperature-41 C
Full
Figure 5-2. Factorial experimental design for dishwasher experiments.
and a water temperature of 41°C. In order to fulfill the factorial requirements, eight experiments
were completed. Additional experiments included replicates and quality assurance tests.
5,3. SOURCE-SPECIFIC METHODOLOGY
A standard protocol was developed for preparing the dishwasher for an experiment. The
following tasks were completed prior to starting an experiment:
• The appropriate dishwasher settings (normal cycle, water heat on or off) were applied
• The dishwasher was started at the beginning cycle (prerinse cycle) and allowed to run to
completion of the first cycle
• The tracer bags were agitated during the first cycle
• The fill of the wash cycle (2nd cycle) was tuned
• For experiments using the water heat option, the dishwasher was allowed to run until the
appropriate elevated temperature (~ 54°C) was reached
• The dishwasher was stopped after fill was complete (or appropriate water temperature was
reached) and the door was opened
• An initial background liquid sample was collected from the dishwasher
• The chemical tracer cocktail was added to the dishwasher basin
• The liquid-phase temperature was recorded
• An initial gas-phase sample was collected and served as the initial gas-phase concentration
5-3
-------
lase
se
• Two liquid-phase samples were collected and averaged, and served as the initial liquid-phase
concentration value.
5.3.1. Sample Schedule
Experiments were designed to last 10 minutes. Experiments using the water heat option
tended to be shorter, depending on cycle time required to reach the appropriate temperature. For
dishwasher experiments, it was expected that a rapid loss of chemical from the liquid pha
would occur in the first minute of operation, followed by a relatively constant liquid-phase
concentration. The liquid-phase sample schedule was designed to reflect this behavior such that
samples were collected at experimental times of 0.25, 0.75, 1.5, 3.0, and 7.0 minutes. Two
additional samples were collected at 10 minutes for applicable experiments. In total, 10 liquid-
phase samples were collected for each dishwasher experiment.
Gas samples were collected for 30 seconds and scheduled such that a liquid sample was
collected at the midpoint of the gas sampling time. At least four gas-phase samples were
collected for each experiment.
5.3.2. Ventilation Rate
A grated exhaust vent was located on the top face of the dishwasher door from which gas
naturally exited the dishwasher. This ventilation rate was estimated using an isobutylene tracer
gas. Isobutylene has a Henry's law constant of 23 m3Iiq/m3gas and will not dissolve appreciably
into dishwasher water. Before starting the dishwasher, isobutylene was introduced at 100 ppm to
the dishwasher headspace. The concentration inside the dishwasher was continuously monitored
using a photo-ionization detector (Photovac™ Microtip). An exponential line was fitted through
data points on a plot of Cg vs. time, with the gas flowrate serving as the adjustable best-fit
parameter.
5.3.3. Parameter Estimation
Experiments were completed using all five chemical tracers (acetone, ethyl acetate,
toluene, ethylbenzene, and cyclohexane). In addition to peaks associated with the five tracers,
the GC method used to analyze liquid-phase samples (see Section 3.4.1) indicated chemical
5-4
-------
peaks associated with compounds in the dishwasher detergent. A compound present in
Cascade™ liquid dishwasher detergent eluted from the GC at the same retention time as ethyl
acetate, thereby masking ethyl acetate's volatilization results. Correcting for this problem by
altering the GC method resulted in a loss of definition for other tracer peaks. Thus, the original
GC method was used to analyze dishwasher experimental samples, and ethyl acetate results were
not reported for this source.
There were two distinct zones of chemical mass transfer for a dishwasher. Within the
first 90 seconds of all experiments, a significant portion of the initial mass was volatilized to the
dishwasher headspace. With the exception of cyclohexane, which completely volatilized, the
dishwasher headspace behaved as if in dynamic equilibrium, a steady-state condition, for the
remainder of the wash cycle (90 seconds to 10 minutes). This phenomenon made it difficult to
estimate values of KLA that were representative of the entire cycle. Thus, values of KLA were
determined for each chemical based on measurements collected within the first 45 seconds of an
experiment. After this time the value of KLA became unimportant because the system had
reached equilibrium conditions; that is, emissions could be determined via a simple equilibrium
analysis without knowledge of specific mass transfer kinetics.
For cyclohexane, Equation 2-19 may be simplified to Equation 5-1 because Cg/Hc for this
compound was negligible compared with C, for initial measured data:
where
Q
t
KL
A
V,
dC, _ KLA
dt
V,
•c,
(5-1)
Chemical concentration in water (M/L3).
Time(T).
Overall mass transfer coefficient (L/T).
Interfacial surface area between water and adjacent air (L2).
Liquid volume (L3).
5-5
-------
For all experiments, cyclohexane was completely stripped from the wash water within 90
seconds. By means of measured liquid-phase concentration values from 0 to 45 seconds, a best-
fit exponential curve (forced through the measured initial liquid-phase concentration value) was
used to estimate KLA for cyclohexane.
Acetone, toluene, and ethylbenzene were not completely transferred from the wash water
because of equilibrium limitations. As with cyclohexane, the maximum volatilization rate for
the other three chemicals occurred within the first 90 seconds. At this point, however, the
system was near equilibrium such that little chemical mass transferred from the liquid phase to
the headspace. As for cyclohexane, an exponential curve was fitted through the first three
liquid-phase concentration data points, and the negative slope of this curve multiplied by the
total liquid volume resulted in a KLA value. For some experiments, this method was less
accurate for toluene, ethylbenzene, and especially acetone, because of the increased Cg/Hc value
as the chemicals approached equilibrium within the headspace. In those cases, the reported
value of KLA would be underestimated.
Using the dishwasher mass balance models (Equations 2-23 and 2-24) also proved
difficult for determining values of KLA based on the initial 45 seconds of operation because of
the nature of gas sampling. First, for several experiments an initial gas-phase sample was not
collected, and thus the initial gas-phase concentration was assumed to be zero. This assumption
maximized the concentration driving force term in Equation 2-19, thus leading to potential
underestimation of KLA. Also, gas-phase samples were collected for 30 seconds, over which
time the average gas-phase concentration was predicted. During the rapid volatilization period
of the first minute, gas-phase concentrations for each chemical increased at an exponential rate,
such that the average measured value did not accurately characterize the headspace
concentration during this time. Later in the experiment, when equilibrium conditions were
reached, the gas samples better represented the actual conditions.
Thus, to be consistent, the method adopted to calculate values of KLA for acetone,
toluene, and ethylbenzene was the same as that used for cyclohexane. Fortunately, the exact
value of KLA for these compounds is not critical because the system reached an equilibrium
condition rapidly for all experiments. Knowing equilibrium will be reached, the amount of mass
5-6
-------
transferred from the liquid phase to the gas phase .can be routinely determined given knowledge
of the headspace ventilation rate and Henry's law constant for a chemical of interest.
5.4. DISHWASHER RESULTS
A total of 11 dishwasher mass transfer experiments and 18 ventilation experiments were
completed to characterize the emission rate from a residential dishwasher. Each mass transfer
and ventilation experiment was completed with the same wash cycle. Dishwasher cycles are
similar in operation, such that experimental results based on a single cycle can be applied to all
cycles in order to predict total chemical emissions during use. The ventilation rates, stripping
efficiencies and mass transfer coefficients (KLA, k,A, kgA, and kg/k,) are presented in this chapter
and are based on the experimental methodology presented in Sections 3.0 and 5.3. In addition,
the effects of liquid temperature, detergent use, and dish loading pattern on each response are
discussed.
The operating conditions for each mass transfer experiment are listed in Table 5-1.
5.4.1. Ventilation Rates
Ventilation rates as well as mass transfer coefficients were difficult to estimate during a
single experiment. Therefore, ventilation rates were determined separately, following the
Table 5-1. Dishwasher experimental operating conditions
Experiment
#
1
2
2 replicate
3
4
4 replicate
5
6
7
8
8 replicate
Liquid
temp.
(°C)
43
42
39
43
45
38
55
55
54
55
53
Liquid
volume
(L)
7.4
7.4
-7.4
7.4
7.4
7.4
7.4
7.4
7.4
7.4
7.4
Headspace
volume
(L)
181
181
181
181
181
181
181
181
181
181
181
Ventilation
rate
(L/min)
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
Cycle
portion
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full
Empty
Full
Full
5-7
-------
methodology given in Section 5.3.2, for operating conditions similar to those used during mass
transfer experiments. A total of 18 ventilation rate experiments were completed, including 11
replicate experiments. A summary of the ventilation experimental operating conditions and
results is provided in Table 5-2.
As shown in Table 5-2, ventilation rates for all combinations of experimental conditions
ranged from 4.6 to 7.2 L/minute. There was little deviation in ventilation rates between different
water temperatures, using detergent or no detergent, and using dishes or no dishes. Thus, all
experimental values were averaged to give an overall ventilation rate of 5.7 L/minute. This
value was applied to all dishwasher mass transfer experimental analyses. The relatively low
ventilation rate of the dishwasher allowed for low chemical emissions during operation and
subsequent accumulation of chemicals in the dishwasher headspace.
A representative data plot for a ventilation experiment is shown in Figure 5-3. The
experimental conditions for this plot were water heat on, detergent present, and full dishwasher
(Ventilation Experiment 18). The slope for the exponential line was -0.0315 with an R2 value of
0.99. Values of R2 ranged from 0.95 to 1.0 for all ventilation plots. These high R2 values
indicated a relatively constant ventilation rate for the duration of the dishwasher cycle. For this
experiment, the washing machine filled at 4.1 L/minute for 99 seconds, resulting in a total liquid
volume of 6.8 L. Given a total volume of 188 L, the remaining headspace volume was 181 L.
The corresponding ventilation rate for this experiment was 181 L multiplied by the negative of
the slope, for a value of 5.7 L/minute.
In addition to the wash cycle, ventilation rates were determined for the entire time of
operation (all cycles used). In general, values based on all of the cycles did not deviate
significantly from the wash cycle results. Thus, the 5.7 L/minute average ventilation rate may be
applied to any dishwasher cycle.
5.4.2. Chemical Stripping Efficiencies
Chemical stripping efficiencies (T|) are reported in Table 5-3 for all tracer chemicals.
Stripping efficiencies for dishwasher experiments were based on the initial and final liquid-phase
5-8
-------
Table 5-2. Dishwasher ventilation rate experimental results
Experiment
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Water heat
on?
No
No
No
No •
No
No
No
No
No
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Detergent
present?
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
Yes
Yes
Yes
Dishes Ventilation rate Exponential fit
present? (L/minute) R2
No
No
No •
No
No '•"
No
No
No
No
No ";
No
No
No
Yes
Yes
Yes
Yes
Yes
6.3
5.0
5
5.9
7.2
4.7
5.6
5.5
5.3
6.5 "
'5.2
4.6
5.5
6.4
5.5
6.2
7.2
5.7
0.95
0.98
0.98
0.98
0.99
0.98
0.99
0.98
0.99
0.95
0.98
0.99
0.99
0.98
0.99
0.98
1.0
099
Time (min)
Figure 5-3. Isobutylene decay because of ventilation for Experiment 18.
5-9
-------
Table 5-3. Chemical stripping efficiencies (r[) for experimental dishwasher
Experiment
#
1
2
2 replicate
3
4
4 replicate
5
6
7
8
8 replicate
Liquid
temperature
(°Q
43
42
39
43
45
38
55
55
54
55
53
Cycle
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-
loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full
Empty
Full
Full
Acetone
r|
(%)
50a
34
45
37
47
42
55a
18
51
37
40a
Toluene
T!
(%)
97a
96
97
96
97
96
98
96
98
97
97a
Ethylbenzene
i\
(%)
97a
97
98
97
98
97
98
97
98
97
98a
Cyclohexane
ri
(%)
100a
100
100
100a
100
100a
100
100a
100
100
100a
'Initial liquid-phase concentration based on average of duplicate samples with a relative difference greater than 20%,
but no more than 36%.
concentrations measured in the basin (Equation 2-2). The time for experiments using the water
heat option was typically 3 minutes shorter than the time for experiments not using this option.
Because the dishwasher headspace reached equilibrium within 2 minutes of operation, additional
chemical volatilization from 7 to 10 minutes was assumed to be minimal. Thus, differences in
experimental times were not accounted for in stripping efficiency results.
Stripping efficiencies for acetone ranged from 18% to 55%, with an overall average
value of 41%. The highest value corresponded to the conditions of a rinse cycle, no dishes, and
water temperature of 55°C.
For acetone, the stripping efficiencies were grouped to complete a factorial main effect
analysis (see Section 3.7 for methodology). To illustrate this analysis, the calculation of the
main effect of dish-loading pattern on stripping efficiency is presented below:
5-10
-------
Corresponding
experiments:
1 — Average (2 and 2 rep)
3 - Average (4 and 4 rep)
5-6
7 - Average (8 and 8 rep)
Difference in
stripping efficiencies
11%
-8.0 %
37 %
12%
Average
13%
As shown in this example calculation, the main effect for dish-loading pattern was 13%.
A positive value indicated that stripping efficiencies for acetone tended to increase with no
dishes present. Acetone's stripping efficiencies were grouped according to dish-loading pattern
and averaged giving values of 48% for an empty dishwasher and 37% for a full dishwasher, both
values similar in magnitude to the overall average.
The second highest main effect on acetone stripping efficiency was detergent use, with a
value of -2.5%. A negative effect indicated that acetone's stripping efficiencies tended to
decrease for wash cycles. Finally, the main effect with water temperature was -2.0%, indicating
that stripping efficiencies for acetone tended to decrease slightly with increasing temperature. In
fact, as expected, stripping efficiencies increased for two of four experiments when water
temperature was increased, and decreased in the other two experiments when water temperature
was increased.
Dishwasher Experiments 2, 4, and 8 were replicated. Through comparison of the acetone
stripping efficiencies for these three experiments, the following relative differences were
calculated: 28% for Experiments 2 and 2 replicate, 11% for Experiments 4 and 4 replicate, and
7.8% for Experiments 8 and 8 replicate.
Stripping efficiencies for toluene ranged from 96% to 98%, with an average value of
97%. Similarly, stripping efficiencies for ethylbenzene ranged from 97% to 98%, also with an
average value of 97%. This narrow range in stripping efficiencies did not allow for the use of a
5-11
-------
factorial analysis, such that no trends related to operating conditions and stripping efficiencies
could be identified. However, stripping efficiencies for toluene and ethylbenzene were
consistently greater than acetone, which has a lower Henry's law constant.
The relative differences in toluene stripping efficiencies between replicate experiments
were 1.0% for Experiments 2 and 2 replicate, 1.0% for Experiments 4 and 4 replicate, and 0%
for Experiments 8 and 8 replicate. Similarly, for ethylbenzene, relative differences were 1.0%
for Experiments 2 and 2 replicate, Experiments 4 and 4 replicate, and Experiments 8 and 8
replicate, respectively.
Finally, 100%of cyclohexane volatilized for every experiment. In fact, 100%
volatilization of cyclohexane was achieved within the first minute of almost all experiments. At
the temperatures listed in Table 5-3, cyclohexane had Henry's law constants between 11
m3]iq/m3gas (Experiment 4 replicate with a temperature of 38°C) and 18 m3liq/m3gas (Experiments 5,
6, and 8 replicate with a temperature of 55°C). Although no operating condition impacts could
be identified, chemical stripping efficiencies did consistently increase for chemicals with
increasing Henry's law constant.
Because cyclohexane achieved 100% volatilization for every experiment, the relative
difference for replicate experiments was consistently 0%.
The chemical stripping efficiency results suggest that, for chemicals with a Henry's law
constant greater than that for toluene, there will be nearly complete removal from the water
whenever a dishwasher is used. This phenomenon is significant for many gasoline constituents,
trichloroethene, teterachloroethene, and radon. For lower volatility chemicals, stripping
efficiency is defined by Henry's law.
5.4.3. KLA Values
Values of KLA for each chemical tracer are reported in Tables 5-4 to 5-7, respectively.
Different tables were used for each chemical to show the factorial main effect analysis on each
5-12
-------
Table 5-4. Acetone K, A values for dishwasher experiments
Experiment
#
1
2
2 replicate
3
4
4 replicate
5
6
7
8
8 replicate
Liquid
temp.
43
42
39
43
45
38
55
55
54
55
53
Cycle
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-
loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full
Empty
Full
Full
KLA
(L/min)
7
4.2
5.8
5.1
6.8
9.4
8.2
1.7
7.6
4.9
5.2
Average =
Dishes
effect3
(L/min)
2
2
-3.0
.1)
6.5
6.5 -
2.5
2.5
2.0
Detergent
effect"
(L/min)
1.9
.1
1.9
.1
0.60
-3.4 -
0.6
3 A
.4
-1.0
Liq. temp.
effect'
(L/min)
1.2
.3
2.5
.0
1.2
--3.3
•-• 2.5
3r\
.u
-0.65
a Dishes effect from full to none.
b Detergent effect from full to none.
c Liquid temperature effect from water heat off to water heat on.
Table 5-5. Toluene K, A values for dishwasher experiments
Experiment
#•'
1
2
2 replicate
3
4
4 replicate
5
6
7 ..
8
8 replicate
Liquid
temp.
43
42
39
43
45
38
55
55
54
55
53
Cycle
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-
loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full ,
•_ Empty
Full
Full
KLA
(L/min)
33
30
32
30
33
34
39
33
38
31
35
Average =
Dishes
effect3
(L/min)
2.0
2
-4.0
.0
6.0
6.0
5.0
5
2.3
Detergent
effect"
(L/min)
3.0
.0
3.0
.0
1
0
1
0
0.25
Liq. temp.
effect6
(L/min)
6 .
2
8
-1.0
6
.. 2
8
.0
-3.8
1 Dishes effect from full to none.
b Detergent effect full to none.
c Liquid temperature effect from water heat off to water heat on.
5-13
-------
Table 5-6. Ethylbenzene KT A values for dishwasher experiments
Experiment
#
1
2
2 replicate
3
4
4 replicate
5
6
7
8
8 replicate
Liquid
temp.
43
42
39
43
45
38
55
55
54
55
53
Cycle
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-
loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full
Empty
Full
Full
KLA
(L/min)
31
32
35
33
35
36
42
36
41
34
37
Average =
Dishes
effect3
(L/min)
-3.0
o n
j.U
-3.0
o n
j.U
6.0
6.0
5.0
5
1.3
Detergent
effect"
(L/min)
-2.0
o n
Z.U
-2.0
o n
z.u
1.0
0
1.0
0
-0.80
Liq. temp.
effect0
(L/min)
11
2
8
0
11
2
8
0
5.3
11 Dishes effect from full to none.
b Detergent effect from full to none.
c Liquid temperature effect from water heat off to water heat on.
Table 5-7. Cyclohexane K, A values for dishwasher experiments
Experiment
#
1
2
2 replicate
3
4
4 replicate
5
6
7
8
8 replicate
Liquid
temp.
43
42
39
43
45
38
55
55
54
55
53
Cycle
type
Rinse
Rinse
Rinse
Wash
Wash
Wash
Rinse
Rinse
Wash
Wash
Wash
Dish-
loading
pattern
Empty
Full
Full
Empty
Full
Full
Empty
Full
Empty
Full
Full
KLA
(L/min)
45
49
58
51
50
62
57
56
50
47
55
Average =
Dishes
effect3
(L/min)
-9.0
•
-5.0
c
J
1.0
1.0
-1.0
1
1
-3.5
Detergent
effectb
(L/min)
-6.0
o
z
-6.0
9
Z
7.0
5.0
7.0
5
1.0
Liq. temp.
effect0
(L/min)
12
2
-1
<
D
12
2
-1
• <
.2
11 Dishes effect from full to none.
b Detergent effect from full to none.
: Liquid temperature effect from water heat off to water heat on.
5-14
-------
combination of operating conditions (see Section 3.7 and 5.4.2 for methodology). The three
factors of the dishwasher two-level factorial arrays were liquid temperature, use of detergent, and
dish-loading pattern. As shown in Tables 5-4 to 5-7, the difference in experimental response was
listed twice, once for each corresponding experiment. Duplicating the listing of each difference
in response, however, does not affect the average value for each variable. The experimental
results for Experiments 2 and 2 replicate, 4 and 4 replicate, and 8 and 8 replicate were averaged,
respectively, before applying factorial analyses.
As shown iri Table 5-4, values of KLA for acetone ranged from 1.7 to 9.5 L/minute, with
an overall average value of 6.0 L/minute The highest value corresponded to the experimental
conditions of a wash cycle with dishes present, and water heat option.
The largest main effect on values of KLA for acetone was 2.0 L/minute for the presence of
dishes. In a manner similar to stripping efficiency results, values of KLA were grouped according
to the presence of dishes resulting in the following average values: 7.0 L/mihute for experiments
using no dishes and 5.5 L/minute for experiments using dishes.
Although a value of KLA was roughly estimated for acetone based on the first three liquid-
phase data points, experimental results clearly show acetone reached an equilibrium condition
within the dishwasher headspace. Figure 5-4 shows the ratios of gas-phase and liquid-phase
concentrations measured during each experiment. The shaded portion of the graph shows the
range of Henry's law constants for acetone calculated based on the correlation given in Section
3.2.1 for all experiments. As shown in Figure 5-4, G/C, values measured at or after 90 seconds
are within or above the shaded region. A possible reason that measured Cg/C, values exceed the
given range of Henry's law constants is the potential inaccuracy of the Henry's law constant for
acetone, at elevated temperatures. Also, experimental error in the liquid standard calibration or
gas standard calibration could lead to higher predictions of Cg/C,. Thus, results suggest that an
assumption of rapid and dynamic equilibrium is valid for lower volatility chemicals, for example,
many disinfection by-products. Unfortunately, Henry's law constants are lacking for many
chemicals at elevated temperatures, a fact that currently hinders accurate predictions of
dishwasher emissions.
5-15
-------
nnifin
0.0120 -
•ts
QJ
u
(A
| 0.0080 -
6
6*
° 0.0040 -
n nnnn .
• . -" •
A : x
•
• X
* @ g
o
a D "
x • »
f 4 ' A
if * +
A i
*
0 50 100 150 200 250 300 350 400 450 500
Time (sec)
Figure 5-4. Comparison of measured Cg/C, to predicted Henry's law constant
for acetone.
Values of KLA for toluene ranged from 30 to 39 L/minute, with an overall average of 33
L/minute. Despite this relatively narrow range of values, a factorial main effect analysis was also
completed for toluene. The results are presented in Table 5-5 for each set of experimental
conditions. The highest main effect was for water temperature, with a value of 3.8 L/minute
Grouping values of KLA according to water temperature and averaging them gave the following
results: 32 L/minute for regular hot water (~ 41 °C) and 35 L/minute for water additionally heated
by a dishwasher heating element (~54°C).
Replicate experiments had the following relative differences between values of KLA for
toluene: 6.5% for Experiments 2 and 2 replicate, 3.0% for Experiments 4 and 4 replicate, and
12% for Experiments 8 and 8 replicate.
Toluene results for Experiment 8 are presented in Figure 5-5. The best-fit KLA value for
this experiment was 31 L/minute. The Henry's law constant for toluene for Experiment 8
(temperature = 55°C) was 0.62 m3Iiq/m3gas. Figure 5-5 illustrates the initial drop in liquid-phase
concentration followed by steady-state conditions. Steady-state conditions were reached because
of the dishwasher ventilation rate. In general, the ratio of Cg/C, for measured data occurring after
5-16
-------
100 seconds was equivalent or slightly greater than the predicted He.nry's law constant for.that
temperature. To further illustrate this approach to equilibrium, the y-axis of Figure 5-5 is shown
magnified in Figure 5-6.
Values of KLA for ethylbenzene were slightly higher than those for toluene, with values
from 31 to 42 L/minute, with an overall average of 36 L/minute. For the temperatures listed in
Table 5-6, ethylbenzene had Henry's law constants ranging from 0.64 m3liq/m3gas and 1.4
m3liq/m3gas compared with 0.40 m3liq/m3gas and 0.62 m3,jq/m3gas for toluene. The factorial main
effects listed in Table 5-6 for ethylbenzene were also similar to those for toluene, with the highest
value being 4.0 L/minute for liquid temperature. Grouping ethylbenzene KLA values according to
liquid temperature resulted in an average value of 34 L/minute for experiments with water heat
off (~ 41°C), and 38 L/minute for water heated by dishwasher heating element (~ 54°C).
Replicate experiments had the following relative differences between values of KLA for
ethylbenzene: 9.0% for Experiments 2 and 2 replicate, 2.8% for Experiments 4 and 4 replicate,
and 8.5% for Experiments 8 and 8 replicate.
Concentration (mg/L]
K> *. ON oo e
i i i i
I
1
1 . • . .•;.,.••-.-..- .
1
-V
\ . - •
-\
- 0 \
0 1—
0
-^^ c^— | 1 . 1 — ^
100 200 300 400
Time (seconds)
o Measured Liquid Values - Liquid Mode
X Measured Gas Values Gas Model Pr
1 Prediction
ediction
500
Figure 5-5. Toluene concentrations for Experiment 8.
5-17
-------
i -
3
£0.8-
.2 0.6-
•O
| 0.2-
U
\
\
• v ... 2* .. «•
-r
"- ™~ •
X
0.0 T 1 II.
0 100 200 300 400
Time (seconds)
o Measured Liquid '
X Measured Gas Va
Values Liquid Model Prediction
lues Gas Model Piedictiou
500
Figure 5-6. Amplification of Figure 5-5 to illustrate approach to equilibrium
conditions for toluene.
Ethylbenzene results for Experiment 8 are plotted in Figure 5-7. This plot is similar to
that of toluene, except the Henry's law constant for ethylbenzene at this temperature is1A
rn^jq/m3^. Thus, at equilibrium, liquid-phase concentrations were less than gas-phase
concentrations; that is, the gas and liquid concentration lines crossed.
Finally, values of KLA for cyclohexane ranged from 45 to 62 L/minute, with an overall
average value of 53 L/minute (see Table 5-7). As expected from its relatively high Henry's law
constant, cyclohexane consistently had the highest KLA value of all tracers for each experiment.
Interestingly, cyclohexane had a slightly larger main effect of-3.5 L/minute associated with dish-
loading pattern compared to 2.0 L/minute for liquid temperature.
Cyclohexane data for Experiment 8 are presented in Figure 5-8, which shows that
cyclohexane has completely volatilized by 60 seconds into the experiment.
Replicate experiments had the following relative differences between values of KLA for
cyclohexane: 17% for Experiments 2 and 2 replicate, 21% for Experiments 4 and 4 replicate, and
16% for Experiments 8 and 8 replicate.
5-18
-------
Concentration (mg/L)
K) *. OS 00 C
o -i
C
1 '
\
~\
-\
-\
*z±. E>_| — . ftr [ 1 r ' :• '" '
) 100 200 300 400 51
Time (sec)
0 Measured Liquid Values Liquid Model Prediction
30
Figure 5-7. Ethylbenzene concentrations for Experiment 8.
The sensitivity of toluene's Henry's law constant on predicted emissions was also
analyzed. The Henry's law constant of toluene would have to be reduced by 70% to reduce the
predicted gas-phase concentrations of Experiment 8 by 10%. Thus, there is a critical value (« 0.2
m3liq/m3gas) above which the accuracy of Henry's law is not as important to the estimation of
chemical emissions from dishwasher use. In this case, even though equilibrium is reached, the
volume of gas is large relative to the volume of liquid such that essentially all of the chemical
mass is transferred to the gas.
5.4.4. Liquid- and Gas-Phase Mass Transfer Coefficients
The extent of chemical mass transfer in a dishwasher is dictated by chemical volatility. .
A chemical with a relatively high Henry's law constant will completely volatilize from the
dishwasher, whereas a chemical with a lower Henry's law constant will be prevented from
completely volatilizing because of equilibrium limitations. For lower volatility compounds,
knowledge of gas-phase resistance to mass transfer is needed only for determining the time
required to reach equilibrium. For higher volatility chemicals, the time to approach complete
stripping is dictated by the liquid-phase mass transfer coefficient.
5-19
-------
Concentration (mg/L)
> l-> tJ M 4X in
i
\
-\
' \
«
A
. ,v
U 1 ' >
0
•«* tt
100 200 300 400 500
Time (sec)
0 Measured Liquid
X Measured Gas Vs
Values
Liquid Model Prediction
Gas Model Prediction
Figure 5-8. Cyclohexane concentrations for Experiment 8.
Because of the rapid approach to equilibrium for all chemicals except cyclohexane, it was
not possible to determine values of kgA during dishwasher experiments. The rate of mass transfer
for cyclohexane is dominated by liquid-phase resistance such that Equation 2.5 may be simplified
to
KLA = k,A (5-2)
Values of k,A for any chemical of interest may be predicted using cyclohexane data and Equation
2.12. The average value of k,A for cyclohexane based on dishwasher experiments was 53
L/minute.
5.4.5. Mass Closure
An important goal for all experiments was to achieve adequate mass closure. For
dishwasher experiments, mass closure was determined for separate experimental periods. An
experimental period was defined when both a liquid-phase sample and a gas-phase sample were
collected. Poor mixing in the initial seconds of a dishwasher experiment tended to lower the
percent mass recovered for each chemical tracer in that mass closure period. Mass closure for
this initial period was also difficult to assess because of gas sampling limitations. For the
remaining three experimental periods, mass closure was consistent for all chemicals and was
5-20
-------
always in the range of 84% to 124%. All mass closure values for dishwasher experiments are
reported in the database in the Appendix.
5-21
-------
-------
6. WASHING MACHINE EXPERIMENTS
In washing machine operation, chemicals originating in a tap water supply can be emitted to
indoor air during the fill and wash/rinse cycles. As previously discussed, the fill cycle is
characterized by different mass transfer mechanisms from those of the wash and rinse cycles,
which are similar in operation. Thus, washing machine volatilization experiments are divided
into two separate groups. Fill cycle experiments are presented in Section 6.1, followed by
wash/rinse cycle experiments in Section 6.2. It should be noted, however, that the results of
these two experimental groups can be combined to determine an overall mass emission rate
during typical washing machine operation.
6.1. FILL CYCLE EXPERIMENTS
6.1.1. Experimental System
A Kenmore™ washing machine (Model No. 25822) was purchased to complete all (both fill
cycle and wash/rinse cycle) washing machine experiments. The experimental washing machine
had a dual basket design with a total interior volume of 150 L (58 cm diameter and 56 cm
height). Operation options included water volume setting (low, medium low, medium, medium
high, high), water temperature setting (cold, warm, hot), agitation speed (slow, fast), and time of
wash cycle (2 to 10 minutes).
The first action of a washing machine is to fill the tub with water. Typically, a washing
machine is directly plumbed to the house water supply. However, for this project, it was
necessary to add chemical tracers to the supply water upstream of the machine. To meet this
need, an auxiliary water supply and pump system was added to provide inlet water to the
machine (see Figure 6-1). A 120 L container served as a tracer reservoir and was filled with 60
to 90 L of tap water (depending on desired fill volume) prior to each experiment. This water was
spiked with the tracer solution in a manner similar to that described in Section 3.2.2. To fill the
washing machine, liquid was pumped at a prescribed flowrate from the tracer reservoir to the
washing machine hose connection using a rotary vane pump (PROCON™) and 1.3 cm OD
Teflon™ tubing. The liquid flowrate was confirmed by timing the collection of a known volume
of liquid. An effort was made to replicate typical washing machine fill rates of 13.1 to 13.8
6-1
-------
1-3 cm
ODX7
Liquid Sample
Tube Port
Liquid Temp.
Thermocouple Teflon Tubing
Port
120 L
Tracer
Reservoir
Water Volume
Setting
Water Temperature
Setting
Degree of Agitation
Liquid Temp.
Thermocouple
Port
Liquid SampL
Tube Port
Rotary
Vane
Setting
Timer]
ial
Gas Temp.
Ther mocouple
Port
Gas Sample
Port
Figure 6-1. Washing machine fill cycle experimental system.
L/minute with the pump and reservoir system. In addition, typical fill times of 3 minutes and 20
seconds for low volume and 6 minutes and 25 seconds for high-volume fills were also used for
appropriate experiments.
For both fill and wash/rinse cycle experiments, the washing machine was configured to allow
for liquid- and gas-phase sampling. A hole 0.32 cm in diameter was drilled in the Washing
machine lid for liquid sampling. During an experiment, 0.32 cm OD Teflon™ tubing was
inserted through the port, and liquid was pumped from the washing machine basin with a
peristaltic pump (Masterflex™, L/S). After the line was flushed for 10 seconds, a liquid sample
was collected hi a 22 mL glass vial as described in Section 3.3.1. For fill cycle experiments, an
additional liquid sample port was drilled in the tracer reservoir lid. Liquid samples from the
tracer reservoir were collected in the same manner as described for the washing machine. Liquid
samples collected from the tracer reservoir represented the initial liquid-phase concentration
6-2
-------
used to solve the fill cycle mass balance equations (Equations 3-8 and 3-9), and were observed to
remain relatively constant during each experiment.
For gas samples, a 0.64 cm ID bore-through stainless steel Swagelok™ fitting was inserted in
the washing machine lid. A 0.64 cm OD sorbent tube was inserted through the fitting into the
washing machine headspace and locked into place with a Teflon™ ferrule located inside the
fitting. A gas sample was pulled through the tube as described in Section 3.3.2, at a sample
flowrate between 0.2 L/minute and 0.4 L/minute. Gas sampling times for wash/rinse cycle
experiments were approximately 30 seconds, whereas a single gas sample was collected for the
duration of a fill cycle experiment.
Liquid-phase temperature was continuously monitored in both the tracer reservoir and the
washing machine. Thermocouple wires were submerged in each basin pool and were connected
to a digital monitor to allow for continuous temperature measurements. There was no significant
difference in temperature between the tracer reservoir liquid and washing machine liquid for the
duration of an experiment.
6.1.2. Experimental Design
Fill cycle experiments were designed to compare the volatilization rate for a standard
condition with the volatilization rate associated with changes in one variable. The fill cycle
standard condition was defined as cold water (T •* 20°C), no detergent, no clothes in machine,
approximately 13.8 L/minute liquid flowrate, low water volume (« 45 L), and a fill time of 3.33
minutes. The independently varied parameters included hot water (T « 50°C), addition of
detergent (« 40 g of Tide detergent), addition of clothes (equivalent liquid volume » 11 L), 8.6
L/minute liquid flowrate (4.75 minute fill time), and high water volume (« 90 L, 6.5 minute fill
time). Six experiments and three replicates were completed.
6.1.3. Source-Specific Methodology
A standard procedure for each fill cycle was developed. Prior to the start of each
experiment the following tasks were completed:
6-3
-------
• The tracer reservoir was filled with at least 60 L of tap water (hot or cold)
• The liquid flowrate was measured and set to the appropriate value
• The tracer cocktail was added to the reservoir water and was mixed manually
• The reservoir tracer solution was mixed for an additional minute
• Detergent and/or clothes were added to the empty washing machine basin when
appropriate
• Two initial liquid samples were collected from the reservoir.
It should be noted that there is no standard protocol for filling a washing machine. Users
commonly add clothes and/or detergent at different times during the filling process, which
incidentally results in the lid being open at different times and for varying time periods. It was
not practical to replicate all possible combinations of procedures associated with loading a
washing machine. Thus, a consistent protocol was adopted for all experiments. The lid was
always closed, and, where applicable, clothes and/or detergent were added to the machine before
the experiment was started.
6.1.3.1. Sample Schedule
Liquid samples were collected from the tracer reservoir throughout the experiment to
monitor any chemical losses, that is, changes in the initial chemical concentrations. Five liquid
samples were collected from the tracer reservoir, and four liquid samples were collected from the
washing machine basin. Liquid samples from the washing machine basin were collected at
experimental tunes of 2.0 and 2.3 minutes. Two additional samples were collected at the end of
filling (3.33 minutes). These liquid sample times were adjusted for longer experiments (low
flowrate and high volume). A single gas sample was collected from the washing machine
headspace for the duration of the experiment, during which time sample volumes were recorded
using a bubble fiowmeter downstream of the adsorbent tube. A final gas sample was also
collected for 30 seconds after experiment completion. Liquid temperatures were monitored for
both the tracer reservoir and the washing machine.
6-4
-------
6.1.3.2. Ventilation Rates
The experimental methodology used to estimate ventilation rates during the fill cycle was
similar to that given in Section 5.3.2. However, the mass balance equation describing the
washing machine headspace during filling incorporated changing liquid and headspace volumes,
as shown:
dt
(6-1)
where
Cg = tracer gas-phase concentration in headspace (M/L3)
Vg = headspace volume (L3)
t = time(T)
Qg = headspace ventilation rate (L3/T)
Cg>in = tracer gas-phase concentration entering headspace (M/L3).
If one assumes the background air was relatively clean (Cgin = 0), Equation 6-1 may be
rewritten as:
where
t
Qg
-dt
dt
(6-2)
tracer gas-phase concentration in headspace (M/L3)
headspace volume (L3)
time(T)
headspace ventilation rate (L3/T).
Further simplifications of Equation 6.2 include rewriting the change in gas volume (dVg/dt)
as - (dV,)/dt, which is equivalent to - Q,. Also, the liquid volume (V,) equals Q,«t Finally, the
gas volume (Vg) may be expressed as the difference between the total washing machine volume
and the liquid volume (Vt - Q^t). The integrated form of Equation 6-2 is then:
6-5
-------
K - * V*M ^fi/1 *-V V t ^M / /y-V y--v \ \ K,0/ ^"V \ t/ >• '
where
Cg = tracer gas-phase concentration in headspace (M/L3)
t = time(T)
Qg = headspace ventilation rate (L3/T).
Q, = liquid flowrate (L3/T)
V, = total machine volume (L3)
Cg>0 = initial tracer gas-phase concentration (M/L3).
The ventilation rate (Qg) was determined by fitting Equation 6-3 to the measured data,
using the procedure outlined in Section 3.6.
6.1.3.3. Parameter Estimation
Ethyl acetate was affected by the presence of detergent. As explained in Section 5.3.3, a
compound present in dishwasher detergent eluted from the GC column at the same residence
time as ethyl acetate, thereby masking ethyl acetate results. Interestingly, a compound present in
Tide™ detergent had an opposite effect on ethyl acetate, because no peak was detected for ethyl
acetate hi experiments involving detergent. This result was replicated with controlled laboratory
experiments in which ethyl acetate was added to vials containing water and detergent.
Apparently, a detergent compound reacted with the ethyl acetate in solution such that ethyl
acetate was no longer measurable using the GC/FID. Thus, ethyl acetate results are not reported
for this cycle.
The duplicate liquid-phase samples collected at the end of the fill cycle were averaged to
determine the C, end value used in Equation 2-2 to estimate chemical stripping efficiencies. If
these duplicate liquid samples were not within 20% of each other, then the average of the
previous liquid samples was used to predict chemical stripping efficiency. The value of C, init in
Equation 2-2 was taken to be the average of liquid-phase concentrations measured in the tracer
reservoir over the course of an experiment.
6-6
-------
As discussed in Section 3.6.2,_mass balance models for the fill cycle could not be solved
analytically, such that a Runge-Kutta second-order numerical solution method was adopted.
This method involved prediction of the following time-dependent parameters: V,, Vg, C,, and Cg,
at 1-second intervals. The value of KLA for each chemical, except acetone, was based on
minimization of the normalized residuals (Equation 3-7) between the liquid-phase concentrations
measured at 2.0, 2.3, and 3.3 (experiment end time) in the washing machine basin and the
model-predicted value at each of these time steps. Because the change in acetone chemical
concentration in the liquid phase was relatively low, the value of KLA for acetone should be
based on gas-phase data. However, for fill cycle experiments, only a single measurement was
collected in the gas phase. Thus, values of KLA for acetone were based on minimizing the
normalized residuals for data in both phases. The normalized residual between the final
measured gas-phase concentration and the final predicted gas-phase concentration in the washing
machine headspace was added to the normalized residuals between the measured liquid-phase
concentrations and model predicted values.
6.1.4. Fill Cycle Results
Nine fill cycle experiments were completed to predict chemical mass emissions. Fourteen
additional experiments were completed to characterize the ventilation rate during the fill cycle.
Fill cycle results can be combined with wash/rinse cycle results presented in Section 6.2.4 to
characterize total mass emissions during typical washing machine use. Based on the
experimental methodology presented in Sections 3.0 and 6.1.3, the ventilation rates, overall
chemical stripping efficiencies, and mass transfer coefficients (KLA, k,A, kgA, and k/k,) are
presented in this chapter. In addition, the effects of liquid temperature, liquid volume, liquid fill
rate, detergent use, presence of clothes, and chemical properties on each response are discussed.
Operating conditions for each mass transfer experiment are listed in Table 6-1. Fill cycle
experiments;were designed to compare a standard condition of cold water, liquid flowrate of
-13.8 L/minute, low liquid volume, no detergent or clothes in the machine, and fill time of 3.33
minutes. Experiments 1 and 1 replicate represented this standard condition. The remaining
experiments have one variable that is different from the standard conditions. The differing
variable is listed in the last column of Table 6-1.
6-7
-------
Table 6-1. Washing machine fill cycle experimental conditions
Experiment
#
1
1 replicate
2
3
4
4 replicate
5
6
6 replicate
Liquid
temp.
(°Q
19
21
19
21
49
47
20
21
19
Fill
time
(min:sec)
3:20
3:20
3:20
3:20
3:20
3:20
6:30
4:45
4:45
Liquid
flowrate
(L/min)
14.6
13.7
13.8
13.7
13.6
13.8
13.7
8.6
8.5
Liquid
final
volume
(L)
49
46
46
46
46
46
89
41
40
Ventilation
rate
(L/min)
55
55
55
55
160
160
55
55
55
Headspace
final volume
(L)
101
104
104
93
104
104
61
109
110
Variable
change
None
None
Detergent
Clothes
Hot water
Hot water
High volume
Low flowrate
Low flowrate
6.1.4.1. Ventilation Rates
Ventilation rates listed in Table 6-1 represent average values based on 14 fill cycle
ventilation rates. The headspace ventilation results listed in Table 6-2 were determined as
explained in Section 6.1.3.2. Several components compose the system ventilation rate. First, the
process of filling involves an expanding liquid pool that naturally displaces air from the washing
machine headspace. The ventilation rate is complicated because additional air is drawn into the
machine by the falling film of water. Also, there are buoyancy effects at elevated temperatures.
As shown in Table 6-2, ventilation rates measured at cold temperatures were lower than at
hot temperatures. Heated water had a significantly higher ventilation rate because of buoyancy
(chimney) effects. Other operating variables (clothes, detergent, high volume, low flowrate) did
not appear to have a significant impact on headspace ventilation. Thus, ventilation rates were
averaged based on liquid temperature. The average cold water ventilation rate was 55 L/minute
and the average hot water ventilation rate was 160 L/minute. These average values were applied
to respective experiments using cold or hot water.
A representative plot for a ventilation experiment is shown in Figure 6-2. The experimental
conditions for this plot were hot water and a liquid flowrate of 13.1 L/minute (Ventilation
Experiment 13). The best-fit ventilation rate for this experiment was 157 L/minute.
6-8
-------
Table 6-2. Washing machine fill cycle ventilation rates
Experiment
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Liquid
temp.
setting
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Cold
Hot
Hot
Fill
time
(min)
3.0
3.25
3.25
3.5
3.5
2.75
3.0
5.5
5.5
6.0
4.75
4.75
2.25
2.0
Liquid
flowrate
(L/min)
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
. 8.6
8.5
13.1
13.1
Ventiliation
rate
(L/min)
49
33
81
33
57
42
47
79
67
52
53
52
157
161
Variable
change
None
None
None
None
None
Clothes
Detergent
High volume
High volume
High volume
Low flowrate
Low flowrate
Hot water
Hot water
6.1.4.2. Chemical Stripping Efficiencies
Chemical stripping efficiencies (r|) for fill cycle experiments are reported in Table 6-3.
Stripping efficiencies for low-volume experiments (Experiments 1 to 4 replicate) were based on
a fill time of 3.33 minutes. Stripping efficiencies for low fill rate experiments (Experiments 6
and 6 replicate) were based on a fill time of 4.75 minutes. Finally, chemical stripping
o.s
1.0 1.5
Time (minutes)
2.0
2.5
Figure 6-2. Isobutylene decay due to ventilation for Experiment 13.
6-9
-------
Table 6-3. Chemical stripping efficiencies (r|) for washing machine fill cycle
Experiment
#
1
1 replicate
2
3
4
4 replicate
5
6
6 replicate
Variable
change
None
None
Detergent
Clothes
Hot water
Hot water
High volume
Low flowrate
Low flowrate
Acetone r|
(%)
2.1
0.96
0.74
3.0
1.2
2.3
1.8
1.2
3.0
Toluene t\
(%)
9.8
13
13
8.2
22
35
17
23
9.7
Ethylbenzene r|
(%)
9.5
13
16
10
20
36
19
24
9.4
Cyclohexane t|
(%)
15
25
26
6.9
28
45
33
37
12
efficiencies for high-volume experiments (Experiment 5) were based on a fill time of 6.5
minutes.
The average stripping efficiencies for the standard condition (liquid flowrate « 13.8
L/minute, low liquid volume, no detergent or clothes in the machine, and fill time of 3.33
minutes) were 1.5% for acetone, 11% for toluene, 11% for ethylbenzene, and 20% for
cyclohexane. In general, stripping efficiencies tended to increase with increasing Henry's law
constant, and toluene and ethylbenzene had similar values for the same experiment. The highest
stripping efficiencies for chemicals (except acetone) were associated with hot water use (average
of Experiments 4 and 4 replicate). The highest stripping efficiency for acetone was for the
condition of clothes in the machine (Experiment 3).
Compared with the standard case, the following conditions led to an increase in chemical
stripping: detergent in the machine for toluene, ethylbenzene, and cyclohexane; clothes in the
machine for acetone; and hot water and low flowrate for all chemicals. In general, however,
overall stripping efficiencies were similar hi magnitude for acetone. An average stripping
efficiency based on all experiments was calculated to be 1.8% for acetone. For the remaining
chemicals, liquid temperature appeared to be a significant factor, resulting in the following
averages: 13% for cold water and 29% for hot water for toluene, 14% for cold water and 28%
for hot water for ethylbenzene, and 22% for cold water and 37% for hot water for cyclohexane.
6-10
-------
Replicate experimental results for the washing machine fill cycles were less consistent than
for other sources. The reasons for high relative differences in replicate experimental results
could not be determined. However, with the exception of one cyclohexane value, the absolute
differences in replicate stripping efficiencies were all within 17%. . .
6.1.4.3. KLA Values
As a washing machine fills, a significant quantity of air is drawn into the underlying pool.
The resulting entrained air influences the rate of chemical volatilization by. increasing a
chemical's gas-phase resistance to mass transfer and by decreasing a chemical's concentration
driving force. These factors are reflected in values of KLA predicted for the fill cycle.
Values of KLA for all chemicals and operating conditions are reported in Table 6-4. Values
of KLA are based on the same fill times discussed for stripping efficiencies. The average values
of KLA for the standard case were 0.23 L/minute for acetone, 2.3 L/minute for toluene, 2.3
L/minute for ethylbenzene, and 4.1 L/minute for cyclohexane. Again, there were general trends
of increasing values of KLA with increasing Henry's law constant as well as similar values for
toluene and ethylbenzene. The impact of entrained air is evident from the 44% difference
between ethylbenzene's KLA and that of cyclohexane for the standard case.
As shown in Table 6-4, there was a great deal of variability in values of KLA for acetone.
Some values could not be determined by the Excel™ solver. This inconsistency likely resulted
from the calculation method of KLA and limited gas-phase data. Thus, a greater emphasis was
placed on the values of KLA for toluene, ethylbenzene, and cyclohexane for fill cycle
experiments. For this particular source, the importance of gas-phase resistance to mass transfer
was evident for these higher volatility compounds.
The highest values of KLA for toluene, ethylbenzene, and cyclohexane were associated with
hot water. The presence of clothes led to a reduction in values of KLA for all chemicals. The
presence of clothes in the washing machine basin visibly reduced the splashing associated with
the falling liquid film and its impact in the underlying pool. In general, experiments completed
with cold water resulted in similar values of KLA. Average values of KLA for cold water
6-11
-------
Table 6-4. Values of KLA for washing machine fill cycles
Experiment
#
1
1 replicate
2
3
4
4 replicate
5
6
6 replicate
Variable
change
None
None
Detergent
Clothes
Hot water
Hot water
High volume
Low flowrate
Low flowrate
Acetone
KLA
(L/min)
0.23
• n/s
n/s
0.086
0.19
0.22
0.038
0.12
1.2
Toluene
KLA
(L/min)
1.8
2.8
4.2
1.5
5.0
8.4
2.5
4.2
3.5
Ethylbenzene
KLA
(L/min)
1.7
2.9
5.0
1.9
4.7
8.4
2.8
4.4
3.7
Cyclohexane
KLA
(L/min)
2.8
5.3
7.5
1.2
5.4
11
4.8
6.4
4.5
Note: Excel solver was unable to find a feasible KLA to fit the model to the measured data.
experiments were 2.9 L/minute for toluene, 3.2 L/minute for ethylbenzene, and 4.6 L/minute for
cyclohexane. For comparison, average values of KLA associated with hot water experiments
were 6.7 L/minute for toluene, 6.6 L/minute for ethylbenzene, and 8.2 L/minute for cyclohexane.
6.1.4.4. Liquid-and Gas-Phase Mass Transfer-Coefficients
Values of KLA for each chemical were separated into the components of k,A and kgA using
Equation 2-5, and a value of kg/lq was determined for each specific experiment. These values are
reported hi Table 6-5. For the fill cycle, values of kg/kj ranged from 4.5 to 20 with an average
value of 9.5 for all experiments. A value of kg/k, was not determined for Experiment 3 because.
the Excel solver could not find a feasible solution for the available data.
Again, the variability associated with values of KLA for acetone prevented them from being
incorporated into the solution matrix. Thus, values reported in Table 6-5 are based solely on
toluene, ethylbenzene, and cyclohexane data. However, the last column of Table 6-5 lists the
predicted average value of KLA for acetone using the reported kg/k] value, Equation 2-15, and
experimental values of KLA for toluene, ethylbenzene, and cyclohexane. By comparison, values
of KLA predicted for acetone in Table 6-5 tend to be lower than those reported for acetone in
Table 6-4. However, values of KLA for acetone for Experiments 4 replicate and 5 are
comparable between the predicted and measured values.
6-12
-------
Table 6-5. Liquid and gas-phase mass transfer coefficients for washing machine fill cycle
experiments
Experiment
#
1
1 replicate
2
3
4
4 replicate
5
6
6 replicate
Chemical
T
EB
C
T
EB
C
T
EB
C
T
EB
C
T
EB
C
T
EB
C
T
EB
C
T
EB
C
T
EB
C
k,A
(L/min)
2.9
2.8
2.9
5.4
5.3
5.5
7.0
8.1
- 7.6
n/s
5.5
4.9
5.4
12
10
11
4.6
5.0
4.9
6.3
6.4
6.5
4.3
4.5
4.5
kgA
(L/min)
21
20
21
25
24
25
47
54
51
n/s
111
101
110
58
50
57
24
26
"25"
54
54
55
80
84
84
Vk.
7.1
4.5
6.7
n/s
20
5.0
5.1
8.5
19
Predicted acetone KLA
(L/min)a
0.022
0.031
0.056
n/s
0.54
0.27
- 0.029
0.066
0.088
aAcetone value of KLA based on kg/k^ Equation 2-15, and values of KLA for toluene, ethylbenzene, and
cyclohexane.
Note: Excel solver unable to find a feasible solution.
6.1.4.5. Mass Closure
Both liquid and gas samples were collected from the filling basin such that the percentage of
mass recovered could be calculated. For fill cycles, the percentage of mass recovered was based
on Equation 3.11 applied for the entire time of fill. The range of mass closure for each chemical
was 96% to 102% for acetone, 90% to 117% for toluene, 84% to 103% for ethylbenzene, and
69% to 102% for cyclohexane. Mass closure values for all experiments are reported in database
in the Appendix.
6-13
-------
6.2. WASH/RINSE CYCLE EXPERIMENTS
6.2.1. Experimental System
The experimental system for wash/rinse cycle experiments was similar to that shown in
Figure 6-1. The same washing machine configured for liquid and gas samples described in
Section 6.1.1 was used, but for wash/rinse cycle experiments it was directly plumbed to the
building water supply. Chemicals were added to the washing machine basin after filling such
that the auxiliary reservoir was not needed. Variable operating conditions for the wash/rinse
cycle included water volume, water temperature, agitation speed, mass of clothing, and presence
of detergent for a wash cycle versus none for the rinse cycle.
The wash/rinse cycle experimental system is shown in Figure 6-3. During the cycle, an
impeller was used to agitate the water. The "normal" wash cycle was used for all experiments.
This cycle can be varied in length. A typical value of 10 minutes was chosen for all
experiments.
Water Temperature
Setting
Degree of Agitation
Setting
/Timer Dial
I/ X
Gas Temp.
Thermocouple
Port
Gas Sample
"~ Port
Figure 6-3. Wash/rinse cycle experimental system.
6-14
-------
6.2.2. Experimental Design
To accommodate all of the variable operating conditions, wash and rinse cycles were
studied using two (2x2x2) factorial arrays as shown in Figure 6-4. The first design consisted
of a wash cycle (« 40 g Tide™ detergent) versus rinse cycle, hot water (T « 50°C) versus cold
water (T * 20°C), and clothes (equivalent liquid volume « 11 L) versus no clothes. The second
array consisted of low water volume (« 45 L) versus high water volume (« 90 L), slow versus
fast agitation speed, and cold water (T « 20°C) versus hot water (T * 50°C). A total of 14
experiments were completed to fulfill both factorial designs, and 3 additional experiments were
completed as replicates.
6.2.3. Source-Specific Methodology .
The following preexperimental tasks were completed for wash/rinse cycle experiments:
The necessary items were added to the washing machine basin (clothes and/or detergent)
The appropriate settings for a particular experiment (water volume, agitation speed, water
temperature) were applied
The washing machine wash time was set to 10 minutes
The washing machine was filled with a known volume of water
The washing machine operation was stopped after the fill was complete (before agitation cycle
began)
_x
Rinse C^\-^
M
Wash (
\J
J
P G.
Factorial #1
(Wash/rinse
^
J
_f
e)
0>
[~XV Clothes
V-L' Vl/ ~ No Clothes
1 1
Cold Hot
Water Water
ow (A
Agitation
TTacf f
L^ ~
J
r /"
• n
Factorial #2
(Wash/rinse
• /^
Agitation ^-^ ^~
Cold H
Water Wa
f
j^
ot
ter
High Volume
~~ Low Volume
Figure 6-4. Wash/rinse cycle factorial experimental design.
6-15
-------
• A background water sample was collected
• The chemical tracer solution was added to the washing machine basin and was mixed well
(manually)
• The washing machine lid was closed
• An initial liquid sample was collected that corresponded to the initial liquid-phase
concentration for an experiment
• An initial gas sample was collected that corresponded to the initial gas-phase concentration
for an experiment.
6.2.3.1. Sample Schedule
A total of 12 liquid samples were collected for each wash/rinse cycle experiment. In
addition to initial samples, liquid samples were collected at the experimental times of .5, 1.25,
1.75,2.75, 3.25, 6.75, and 7.25 minutes. Two additional samples were collected at 10 minutes.
These sampling tunes corresponded to the start and end times of each respective gas sample. For
example, a gas sample was collected from time 0 to 30 seconds, 1.25 to 1.75, and so on.
Including the initial sample, six gas samples were collected for each experiment. Liquid and
gas-phase temperatures were recorded for the duration of the experiment.
6.2.3.2. Ventilation Rates
Washing machines are characterized by a relatively high ventilation rate. This rate was
determined for all wash/rinse cycle experimental conditions using the same methodology as
described hi Section 5.3.2. Ventilation rates determined using isobutylene decay were used in
wash/rinse cycle mass balance models with data from mass transfer experiments.
6.2.3.3. Parameter Estimation
An important measurement used to determine chemical stripping efficiencies and mass
transfer coefficients was the initial liquid-phase concentration. For several experiments, the
liquid-phase concentration increased in magnitude for various lengths of time before decreasing
as expected. This initial increase was likely caused by improved mixing of the chemical tracer
solution in the washbasin. For consistency, each chemical's stripping efficiency was calculated
based on the highest measured liquid-phase concentration during an experiment and the final
6-16
-------
measured liquid-phase concentration. This procedure resulted in experimental stripping
efficiencies based on different time periods; for example, an experiment with the highest liquid-
phase concentration at time zero had a total time of 10 minutes, and an experiment with the
highest value occurring after 2 minutes into the experiment had a total time of only 8 minutes.
To correct for this time difference, a plot was constructed based on measured liquid-phase
concentration values versus time. For experiments with a late initial concentration peak, a curve
was fitted to the data and extended to reach 10 minutes. On the basis of the graph's liquid-phase
concentration value at 10 minutes and the measured initial concentration, a 10-minute stripping
efficiency was reported for every experiment.
Values of KLA for each chemical were calculated based on measurements collected from an
experimental time of 180 seconds to the end of the experiment. This method ensured that the
washing machine contents were well mixed. The difference in experimental time should not
affect the reported KLA values for each chemical, as long as equilibrium conditions did not exist
in the machine's headspace. Values of KLA for acetone and ethyl acetate were based on
minimizing the residuals between the model and gas-phase data. Values of KLA for toluene,
ethylbenzene, and cyclohexane were based on minimizing the residuals between the model and
liquid-phase data. For experiments with conditions leading to relatively high volatilization rates,
the more volatile chemicals often had results below the predetermined method detection level
(see Section 3.5.4). In these cases, the determination of KLA was modified to include only
measurements meeting this quality assurance requirement, that is, above method detection limit.
6.2.4. Wash/Rinse Cycle Results
A total of 17 wash/rinse cycle mass transfer experiments and 17 ventilation experiments
were completed to characterize the emission rate from a residential washing machine during
these cycles. Wash and rinse cycle results can be combined with fill cycle results presented in
Section 6.1.4 to characterize total mass emissions during typical washing machine use. Based on
the experimental methodology presented in Sections 3.0 and 6.2.3, the ventilation rates, overall
chemical stripping efficiencies and mass transfer coefficients (KLA, k^A, kgA, and kg/k,) are
6-17
-------
presented in this chapter. In addition, the effects of liquid temperature, liquid volume, detergent
use, mass of clothes, agitation speed, and chemical properties on each response are discussed.
The operating conditions for each mass transfer experiment are given in Table 6-6.
6.2.4.1. Ventilation Rates
It was difficult to estimate ventilation rates and mass transfer coefficients during a single
experiment. Therefore, ventilation rates were predicted separately, following the methodology
given in Section 5.3.2, for similar operating conditions used during mass transfer experiments.
A total of 17 ventilation rate experiments were completed including 9 replicate experiments. A
summary of the ventilation experimental operating conditions and results is provided in Table
6-7.
As shown in Table 6-7, ventilation rates measured at cold temperatures were significantly
lower than ventilation rates measured at hot temperatures. The heated water led to a buoyancy
(chimney) effect, which acted to flush the headspace at a faster rate. Other factors such as
agitation speed, mass of clothing, presence of detergent, and volume of water had less impact on
Table 6-6. Washing machine wash/rinse cycle experimental operating conditions
Experiment
#
1,A.
1, A replicate
2,B
3
3 replicate
4
5
6
7
8
C
C replicate
D
B
F
G
H
Liquid
temp.
(°C)
24
22
49
23
22
51
21
50
18
49
21
21
51
20
49
18
50
Liquid
volume
(L)
47
49
48
49
47
49
50
47
49
49
82
95
96
48
49
95
94
Headspace
volume
(L)
103
101
102
101
103
101
88
92
90
90
58
55
54
102
101
55
56
Ventilation
rate
(L/min)
53
53
200
53
53
200
53
200
53
200
53
53
200
53
200
53
200
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Detergent
present?
No
No
No
Yes
Yes
Yes
NO-
NO
Yes
Yes
No
No
No
No
No
No
No
Clothes
present?
No
No
No
No
No
No
Yes
Yes
Yes
Yes
No
No
* No
No
No
No
No
6-18
-------
Table 6-7. Ventilation rate experiment results
Experiment
#
1
2
3
4
5'
6
7
8
9 '
10
11
12
13
14
15
16
17
Water
temperature
Cold
Cold
Cold
Cold
Cold
Cold
Cold
. Cold
Cold
Cold
Cold
Hot
Hot
Hot
Hot
Hot
Hot
Water
volume
Low
Low
Low
High
High
Low
Low
Low
Low
Low
Low
Low
Low
Low
High
Low
Low
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Slow
Detergent
present?
No
No
No
No
No
No
No
•No.
Yes
Yes
No
No
No
No
No
No
No
Clothes
present?
No
No
No
No
No
No
No
No
No
No
Yes .
No
No
No
No-
Yes
Yes
Ventilation
rate (L/min)
50
63
43
35
38
78
41
51
41
64
77
116
254
160
246
184
210
the wash/rinse cycle ventilation rate. To determine an appropriate ventilation rate to use in
conjunction with mass transfer data, ventilation experimental values were grouped according to
water temperature. The average cold water ventilation rate was assumed to be 53 L/minute and
was applied to all mass transfer data analyses based on experiments using cold water. The
average hot water ventilation rate was assumed to be 200 L/minute and was applied to all mass
transfer data analyses based on hot water experiments.
A representative data plot for a ventilation experiment is shown in Figure 6-5. The
experimental conditions for this plot were cold water, no clothes, no detergent, low water
volume, and fast agitation. The slope for the exponential line was -0.492 with an R2 value of
0.99. Values of R2 ranged from 0.88 to 0.997 for all ventilation plots, with all but one value
above 0.93. These high correlation values indicated a relatively constant ventilation rate for the
duration of the wash/rinse cycle. For this experiment, the washing machine filled at 13.8
L/minute for 3.43 minutes, resulting in a total liquid volume of 47 L. Based on a total volume of
150 L, the remaining headspace volume was 103 L. The corresponding ventilation rate for this
experiment was 103 L multiplied by the negative of the slope for a value of 51 L/minute.
6-19
-------
6.2.4.2. Chemical Stripping Efficiencies
Chemical stripping efficiencies are reported in Tables 6-8 to 6-16 for each chemical,
respectively. The results for each chemical are reported in two tables based on each factorial
design. The three factors incorporated into the first group were liquid temperature, mass of
0.5
i.o
1.5
2.0 2.5
Time (min)
3.0
3.5
4.0
4.5
Figure 6-5. Isobutylene decay due to ventilation for Experiment 8.
Table 6-8. Acetone stripping efficiencies for washing machine wash/rinse cycle —
Factorial^
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
9
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
9
no
no
no
no
no
no
Yes
Yes
Yes
Yes
Stripping
efficiency
(%)
7.1
15
36
7.0
5.1
30
19
9.4
20
22
Average =
Clothes
effect3 (%)
8 ft
27
8 0
8.0
-8.0
27
-8.0
8.0
4.7
Detergent
effect" (%)
1 0
6.0
1 fl
6.0
-1.0
-13
-1.0
-13
-2.2
Liquid
temperature
effect0 (%)
75
25
1 8
18
-9.6
-9.6
2.0
2.0
8.9
"Clothes effect from full to none.
""Detergent effect from 40 grams to none.
"Liquid temperature effect from cold to hot.
6-20
-------
Table 6-9. Acetone stripping efficiencies for washing machine wash/rinse cycle
Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Stripping
efficiency
(%)
7.1
15
36
3.4
4.8
3.1
16
31
10
15
Agitation
speed
effect8 (%)
-5.0
5.0
.9
-12
-5.0
5.0
-5.9
-12
Liquid
volume
effect" (%)
6.9
33
6.9
33
6.0
16
6.0
16
Liquid
temperature
effect0 (%)
25
25
1.0
1.0
15
15
5.0
5.0
Average = -4.5 15 11
bLiquid volume effect from high to low.
°Liquid temperature effect from cold to hot.
Table 6-10. Ethyl acetate stripping efficiencies for washing machine wash/rinse cycl
Factorial #2
Experiment
#
. A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Stripping
efficiency
(%)
12
8.1
48
5.2
5.2
5.1
16
34
7.8
22
Agitation
speed
effect3 (%)
.0
14
.6
-17 '-
-6.0
14
-2.6
-17
Liquid
volume
effect" (%)
4.8
43
4.8
43
8.2
12
8.2
12
Liquid
temperature
effect0 (%)
38 ":"
38
0.10
0.10
18
18
14
14
Average= -2.9 17 18
aAgitation speed effect from fast to slow.
"Liquid volume effect from high to low.
"Liquid temperature effect from cold to hot.
6-21
-------
Table 6-11. Toluene stripping efficiencies for washing machine wash/rinse cycle
Factorial#l
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
9
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
9
no
no
no
no
no
no
Yes
Yes
Yes
Yes
Stripping
efficiency
(%)
72
65
95
33
34
67
45
56
42
62
Average =
Clothes
effect* (%)
24
39
8 A
5.0
24
39
-8.0
5.0
15
Detergent
effect" (%)
35
28
35
28
3.0
-6.0
3.0
-6.0
15
Liquid
temperature
effect0 (%)
26
26
33
33
11
11
20
20
23
"Clothes effect from full to none.
""Detergent effect from 40 grams to none.
liquid temperature effect from cold to hot.
Table 6-12. Toluene stripping efficiencies for washing machine wash/rinse cyck
Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Stripping
efficiency
(%)
72
65
95
26
28
33
70
99
24
33
Agitation
speed
effect3 (%)
1 0
-4.0
30
0.0
-1.0
-4.0
3.0
0.0
Liquid
volume
effect" (%)
42
62
42
62
46
66
46
66
Liquid
temperature
effect0 (%)
26
26
6.0
6.0
29
29
9.0
9.0
Average = -0.50 54 18
"Agitation speed effect from fast to slow.
bLiquid volume effect from high to low.
•Liquid temperature effect from cold to hot.
6-22
-------
Table 6-13. Ethylbenzene stripping efficiencies for washing machine wash/rinse cycle
Factorial#l . .
Experient
#
1
1 replicate
2
3
3 replicate
4
• 5
6
7
8
Liquid
Temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
?
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
9
No
No
No
No
. no
No
Yes
Yes
Yes
Yes
Stripping
efficiency
(%)
76
69
97
36
37
72
57
65
54
69
Clothes
effect" (%)
16
32
• • -17
3.0
16
32
-17
3.0
Detergent
effect" (%)
36
25
36
25
3.0
-4.0
3.0
-4.0
Average = 8.5 15
Liquid
temperature
effect0 (%)
24
24
35
35
8.0
8.0
15
15
21
"Detergent effect from 40 grams to none.
°Liquid temperature effect from cold to hot.
Table 6-14. Ethylbenzene stripping efficiencies for washing machine wash/rinse cych
Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Stripping
efficiency
(%)
76
69
97
28
31
32
74
99
24
34
Agitation
speed
effect3 (%)
1.0
-2.0
6.0
-2.0
-1.0
-2.0
6.0
-2.0
Liquid
volume
effect" (%)
43
65
43
65
50
65
50
65
Average = 0.25 56
Liquid
temperature
effect0 (%)
24
24
2.0
2.0
25
25
10
10
15
"Liquid volume effect from high to low.
°Liquid temperature effect from cold to hot.
6-23
-------
Table 6-15. Cyclohexane stripping efficiencies for washing machine wash/rinse cyck
Factorial#l
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
9
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
9
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Stripping
efficiency
(%)
99
99
100
82
76
98
79
84
79
94
Clothes
effect" (%)
20
16
0.0
4.0
20
16
0.0
4.0
Detergent
effect" (%)
20
2.0
20
2.0
0.0
-10
0.0
-10
Liquid
temperature
effect' (%)
1.0
1.0
19
19
5.0
5.0
15
15
Average = 10 3.0 10
•Clothes effect from full to none.
"Detergent effect from 40 grams to none.
*Liquid temperature effect from cold to hot.
Table 6-16. Cyclohexane stripping efficiencies for washing machine wash/rinse cycl
Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
Stripping
efficiency
(%)
99
99
100
36
44
44
100
100
48
62
Average =
Agitation
speed
effect3 (%)
10
0.0
Q n
-18
-1.0
0.0
-8.0
-18
-6.8
Liquid
volume
effect" (%)
59
56
59
56
52
38
52
38
51
Liquid
temperature
effect0 (%)
1 0
1.0
40
4.0
0.0
0.0
14
14
4.8
"Agitation speed effect from fast to slow.
"Liquid volume effect from high to low.
"Liquid temperature effect from cold to hot.
detergent, and mass of clothes. The second group involved an investigation of other factors:
liquid temperature, liquid volume, and agitation speed. In order to focus on single-variable
effects, detergent and clothes were not used for tibis second group of experiments.
6-24
-------
For each group, the results of the factorial main effect analysis (see Section 3.7 for
methodology) are given. To illustrate this analysis, the calculation of the main effect of
detergent on acetone's stripping efficiency in factorial #1 is shown below.
Corresponding
Experiments:
Average (1 and 1 rep) to Average (3 and 3 rep) =
2 to 4
5 to 7
6 to 8
Difference in
Stripping Efficiencies
-1.0%
6.0 %
-1.0%
-13%
Average
-2.2 %
As shown in Table 6-8, the difference in experimental response was listed twice, once for
each corresponding experiment. Replicating the listing of each response, however, does not
affect the average value for each variable. As shown in the example, the results for Experiments
1 and 1 replicate, and Experiments 3 and 3 replicate were averaged, respectively, before
applying any factorial analyses. Tables 6-9 to 6-16 follow this same format.
Acetone stripping efficiencies are reported for each factorial group in Tables 6-8 and 6-9.
For both groups of factorials, stripping efficiencies for acetone ranged from 3.1% to 36%. The
highest stripping efficiency value was for the conditions of low water volume, no clothes or
detergent, hot water, and slow agitation. The second highest value associated with the second
factorial group was 31%, also associated with hot water, no clothes or detergent present, and low
water volume. However, this value occurred during a fast agitation speed. It was expected that
for similar operating conditions, experiments completed at a higher temperature would result in
higher stripping efficiencies because of the corresponding increase in Henry's law constant. For
the temperatures listed in Table 6-6, Henry's law constants for acetone ranged from 0.00085
m3Iiq/m3gas (Experiments 7 and G at 18°C) to 0.0051 m3Iiq/m3gas (Experiments 4 and D at 51°C).
The first factorial analysis for acetone stripping efficiencies was based on values calculated
using Experiments 1 through 8. In keeping with these values, the highest main effect was 8.9%
for the single variable factor of liquid temperature. The main effect from differences in liquid
temperature was calculated by subtracting cold water stripping efficiencies from corresponding
(similar amounts of clothing and detergent present) hot water stripping efficiencies. A positive
6-25
-------
effect indicated an absolute increase in stripping efficiency with increasing water temperature.
This result was expected, based on the increasing Henry's law constant as described above.
When the experiments were grouped according to liquid temperature and the respective stripping
efficiencies averaged, the folio whig values resulted: 12% for cold water experiments
(Experiments 1,1 replicate, 3, 3 replicate, 5, and 7) and 24% for hot water experiments
(Experiments 2,4, 6, and 8).
A more practical way to group the experimental results was to combine the liquid
temperature effects with using clothes in a wash or rinse (no detergent present) cycle. The
average stripping efficiencies were 20% and 19% for cold water use during wash and rinse
cycles, respectively, and 22% and 9.4% for hot water use during wash and rinse cycles,
respectively.
The second factorial group also included liquid temperature as a factor (11% main effect).
However, liquid volume had a slightly greater main effect, with a value of 15%. The main effect
from differences hi liquid volume was calculated by subtracting high water volume stripping
efficiencies from low water volume stripping efficiencies. Thus, a positive 14% indicated an
absolute increase in stripping efficiency with decreasing water volume. At lower water volumes,
the total kinetic energy (TKE) resulting from agitation of the water surface increases, thereby
increasing the potential for chemical volatilization.
When the second factorial results were grouped according to liquid volume, the following
average stripping efficiencies resulted: 21% for low volume experiments and 7.3% for high
volume experiments. Liquid temperature also had a significant impact on acetone stripping
efficiencies. Grouping experiments according to volume and liquid temperature resulted in the
following average values: 13% for low volume and cold water experiments, 34% for low
volume and hot water experiments, 6.1% for high volume and cold water experiments, and 9.1%
for high volume and hot water experiments.
As for all chemicals, the reported acetone stripping efficiencies represent a range of possible
transfer efficiencies for different operating conditions. A better estimation of chemical
6-26
-------
volatilization may be made using KLA values reported in Section 6.2.4.3. These values were
based on a well-mixed initial liquid-phase concentration, rather than the highest peak.
Washing machine wash/rinse cycle Experiments 1 (A), 3, and C were replicated. When the
acetone stripping efficiencies for these three experiment were compared, the following relative
differences were calculated: 71 % for Experiments 1 (A) and 1 (A) replicate^ 31 % for
Experiments 3 and 3 replicate, and 34% for Experiments C and C replicate.
Because of detergent interaction discussed in Section 6.2.3, only ethyl acetate results for the
second factorial group are reported in this section. As shown in Table 6-10, ethyl acetate
stripping efficiencies ranged from 5.1% to 48%. Again, the highest stripping efficiency
corresponded to the conditions of low water volume, low agitation speed, and hot water. The
highest main effect for ethyl acetate stripping efficiencies was liquid temperature, with a value of
18%. Grouping the stripping efficiencies according to liquid temperature, resulted in a cold
water average of 9.1% and a hot water average of 27%. For the temperatures listed in Table 6-6,
Henry's law constants for ethyl acetate ranged from 0.0037 m3Iiq/m3gas to 0:016 m3liq/m3gas.
The second highest factor on ethyl acetate stripping efficiencies was liquid volume, with a
value of 17%. As with acetone, the stripping efficiencies for ethyl acetate may be grouped
according to liquid volume and liquid temperature such that 12% is the average for cold water
and low volume, 41% is the average for hot water and low volume, 6.1% is the average for cold
water and high volume, and 14% is the average value for hot water and high volume.
Replicate experiments with ethyl acetate results included Experiments A and A replicate
and C and C replicate. Stripping efficiencies were within 39% for Experiments A and A
replicate and were identical for Experiments C and C replicate.
Toluene stripping efficiencies ranged from 24% to 99% for both factorial experimental
groups (Tables 6-11 and 6-12). The highest stripping efficiency corresponded to conditions of
hot water, low volume, no clothes or detergent present, and fast agitation. Again, hot water led
6-27
-------
to higher stripping efficiencies. For temperatures listed in Table 6-6, Henry's law constants for
toluene ranged from 0.22 m3Iiq/m3gas to 0.57 m3liq/m;
gas-
Toluene stripping efficiencies exhibited a wide range of values depending on associated
operating conditions. Thus, the factorial analysis was a useful tool in determining variable
impacts. For the first factorial group, the variable with the single highest effect was liquid
temperature at a value of 23%. Grouping stripping efficiencies according to liquid temperature
resulted hi an average value of 49% for cold water experiments and 70% for hot water
experiments.
The clothes main effect was 15%, indicating that stripping efficiencies tended to decrease
with clothes hi the machine. This phenomenon was previously observed by Shepherd et al.
(1996) for chloroform in washing machines, and is likely caused by suppression of turbulent
kinetic energy by clothes hi the washbasin. The cold water wash and rinse cycles with clothes
had stripping efficiencies of 42% and 45%, respectively. The hot water wash and rinse cycles
with clothes were characterized by higher stripping efficiencies of 62% and 56%, respectively.
Both the cold water and hot water wash and rinse cycles had lower stripping efficiencies
than the averages calculated based on temperature. This difference may be attributed to the
impact of detergent and clothes on stripping efficiencies. The detergent main effect was also
15%, indicating that stripping efficiencies tended to decrease for wash cycles. Surfactants
present hi detergent act to suppress chemical volatilization by increasing liquid-phase resistance
to mass transfer. Thus, it is not coincidental that the presence of detergents has a greater effect
on those tracers that were dominated by liquid-phase resistance to mass transfer (toluene,
ethylbenzene, cyclohexane) than those dominated by gas-phase resistance to mass transfer
(acetone).
The second factorial group was used to investigate the impacts of water temperature, water
volume, and agitation speed. A wide range of values also characterizes this group of experiment
results. For this group, the effects of liquid volume far exceeded the effects of temperature and
agitation speed, with a value of 54%. Grouping experimental stripping efficiencies according to
6-28
-------
liquid volumes resulted in an 80% average for low-volume experiments, and 29% average for
high-volume experiments. Accounting for the second highest factor of liquid temperature
further separated these averages. The average stripping efficiency for low volume and cold
water was 69%, the average for low volume and hot water was 97%, the average for high
volume and cold water was 26%, and the average for high volume and hot water was 33%. As a
worst case scenario, operating at conditions of hot water and low water volume, virtually all of
the toluene mass initially present in the washing machine basin would be emitted to room air.
However, operating with conditions of high water volume with cold water, only 25%, of the
toluene mass would be emitted. Thus, using a 100% volatilization estimate would dramatically
overestimate chemical emissions for several operating conditions.
Replicate experiment results for toluene had relative differences of 10% for Experiments
1(A) and 1(A) replicate, 3.0% for Experiments 3 and 3 replicate, and 7.4% for Experiments C
and C replicate.
As discussed in Section 3.2.1, toluene and ethylbenzene have similar Henry's law constants
and thus should yield similar volatilization results. As shown in Tables 6-13 and 6-14,
ethylbenzene stripping efficiencies ranged from 24% to 99%. This range was similar in
magnitude to the range of stripping efficiencies reported for toluene. Over 1.7 experiments, the
average relative difference between toluene and ethylbenzene stripping efficiencies was 8.3%.
Main effect values for ethylbenzene were only slightly different from those for toluene.
Again, for the first factorial group, liquid temperature had the dominant main effect on stripping
efficiency, with a value of 21%. Contrary to results obtained for toluene, there was a difference
in the magnitude of the main effect associated with clothes and detergent. In fact, detergent had
a main effect value almost twice as high as that observed for clothes. Thus, there was a greater
difference between wash and rinse cycles for this compound. However, ethylbenzene stripping
efficiencies were similar for wash and rinse cycles at similar temperatures.
For the second factorial group, ethylbenzene again shared common main effects with
toluene. For example, the main effect for liquid volume was 56% and by far exceeded other
6-29
-------
main effect values. Grouping stripping efficiencies according to this one effect resulted in an
average stripping efficiency of 83% for low liquid volume and 30% for high liquid volume,
again a factor of three difference. Adding temperature effects to these averages resulted in
values of 73% for low volume and cold water, 98% for low volume and hot water, 28% for high
volume and cold water, and 33% for high volume and hot water.
Replicate experiment results for ethylbenzene stripping efficiencies were 10% for
Experiments 1(A) and 1(A) replicate, 2.7% for Experiments 3 and 3 replicate, and 10% for
Experiments C and C replicate.
Finally, cyclohexane stripping efficiencies ranged from 36% to 100% (see Tables 6-15 and
6-16). For similar experimental conditions, cyclohexane consistently had the highest stripping
efficiency of the five experimental tracers. Experiments involving hot or cold water, fast or slow
agitation, and low liquid volume resulted in stripping efficiencies of at least 99%. For the
temperatures listed in Table 6-6, Henry's law constants for cyclohexane ranged from 5.8
m3liq/m3glisto 16m3Iiq/m3gas.
Presence of clothes in the machine and water temperature had equal main effect magnitudes
for cyclohexane in the first factorial group. Grouping cyclohexane stripping efficiencies
according to these two factors resulted in the following averages: 89% for no clothes and cold
water, 99% for no clothes and hot water, 79% for clothes and cold water, and 89% for clothes
and hot water. Washing and rinsing clothes in cold water each led to a stripping efficiency of
79%. A stripping efficiency of 89% was observed for wash and rinse cycles involving clothes
and hot water.
For factorial group #2, cyclohexane had a wider range of experimental results. This wider
range derives primarily from the large main effect value for liquid volume. This effect was
approximately seven tunes greater than the main effects for the other two variables. Grouping
stripping efficiencies according to liquid volume resulted in an average value of 100% for low-
volume experiments and 45% for high-volume experiments.
6-30
-------
Replicate experiments had the following relative differences in results: 0% for Experiments
1(A) and 1(A) replicate, 7.6% for Experiments 3 and 3 replicate, and 20% for C and C replicate.
In general, the presence of clothes and/or detergent and using high water volumes resulted
in reduced chemical stripping efficiencies. Accounting for these variable effects leads to
significantly lower transfer efficiencies than the often assumed value of 100%.
6.2.4.3. KLA Values
Values of KLA for each chemical tracer are reported in Tables 6-17 to 6-25, using the same
two factorial groups as for chemical stripping efficiencies. Again, the first factorial group was
designed to investigate the effects of liquid temperature, use of detergent, and presence of
clothes on KLA. The second factorial group was designed to investigate the effects of liquid
temperature, liquid volume, and agitation speed on KLA. Values of KLA for acetone and ethyl
acetate were based on minimizing the residuals between the measured and predicted gas-phase
data (see Section 3.6.2 for methodology). Values of KLA for the remaining tracers were based
on minimizing the residuals between the measured and predicted liquid-phase data, tables 6-17
through 6-25 have a similar format to that of Tables 6-8 to 6-16, except that the main effects are
based on values of KL A.
Values ofKLAfor acetone spanned nearly two orders of magnitude, ranging from 0.0075 to
0.31 L/minute (see Tables 6-17 and 6-18). The highest value corresponded to the experimental
conditions of hot water, low water volume, no detergent or clothes present, and fast agitation.
The highest value in the first factorial also corresponded to conditions of hot water, low water
volume, no detergent or clothes present, but slow agitation.
The largest main effect for the first factorial group was liquid temperature, with a value of
0.10 L/minute. In a manner similar to that for stripping efficiency results, values of KLA were
grouped according to liquid temperature, resulting in the following average values: 0.024
L/minute for cold water experiments and 0.13 L/minute for hot water experiments.
6-31
-------
Table 6-17. Acetone KLA values for washing machine wash/rinse cycle—Factorial #1
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
?
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
?
no
no
no
no
no
no
Yes
Yes
Yes
Yes
KLA
(L/min)
0.069
0.024
0.30
0.011
0.0084
0.022
0.024
0.099
0.0075
0.082
Clothes
effect"
(L/min)
0.023
0.20
0.0022
-0.060
0.023
0.20
0.0022
-0.060
Detergent
effectb
(L/min)
0.037
0.28
0.037
0.28
0.017
0.017
0.017
0.017
Liq. temp.
effect0
(L/min)
0.25
0.25
0.012
0.012
0.075
0.075
0.072
0.072
Average = 0.042 0.087 0.10
"Clothes effect from full to none.
bDetergent effect from 40 grams to none.
°Liquid temperature effect from cold to hot.
Table 6-18. Acetone KLA values for washing machine wash/rinse cycle—Factorial #2
Experiment.
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
Temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
KLA
(L/min)
0.069
0.024
0.30
0.024
0.020
0.15
0.048
0.31
0.023
0.086
Agitation
effect"
(L/min)
Onni n
.UUJ.U
-0.010
0.0010
0.064
-0.0010
-0.010
0.0010
0.064
Liq. volume
effectb
(L/min)
0.025
0.15
0.025
0.15
0.025
0.22
0.025
0.22
Liq. temp.
effect0
(L/min)
0.25
0.25
0.13
0.13
0.26
0.26
0.063
0.063
Average = 0.013 0.11 0.18
"Agitation speed effect from fast to slow.
bLiquid volume effect from high to low.
°Liquid temperature effect from cold to hot.
6-32
-------
Table 6-19. Ethyl acetate KLA values for washing machine wash/rinse cycle—Factorial #2
Experiment.
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
KLA
(L/min)
0.15
0.073 V
0.61
0.053
0.039
0.25
0.091
0.82
0.055
0.13
Agitation
effect8
(L/min)
0.019
-0.21
OnnoA
.uuyu
0.12
0.019
-0.21
-0.0090
0.12
Liq. volume
effect"
(L/min)
"0.064
-•• 0.36
0.064
0.36
, 0.036
0.69
0.036
0.69
Liq. temp.
effect0
(L/min)
0.50
0.50
0.20
0.20
, 0.73
:• 0.73
0.075
0.075
Average = -0.020 0.29 0.38
"Agitation speed effect from fast to slow.
bLiquid volume effect from high to low.
"Liquid temperature effect from cold to hot.
Table 6-20. Toluene KLA values for washing machine wash/rinse cycle—Factorial #1
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
?
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
9
no
no
no
no
no
no
Yes
Yes
Yes
Yes
KLA
(L/min)
9.4
7.1
15
1.5
2.5
3.5
0.84
3.9
0.58
2.1
Clothes
effect3
(L/min)
: 7.5.
11
1.4
1.4
7.5
11
1.4
1.4
Detergent
effectb
(L/min)
6.3
12
6.3
12
0.26
1.8
0.26
1.8
Liq. temp.
effect0
(L/min)
6.7
6.7
1.5
1.5
3.1
3.1
1.5
1.5
Average = 5.3 5.0 3.2
"Clothes effect from full to none.
'Detergent effect from 40 grams to none.
cLiquid temperature effect from cold to hot.
6-33
-------
Table 6-21. Toluene KLA values for washing machine wash/rinse cycle—Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
KLA
(L/min)
9.4
7.1
15
2.7
2.9
3.3
11
38
1.5
1.5
Agitation
effect3
(L/min)
27
. 1
-23
1.3
1.8
-2.7
-23
1.3
1.8
Liq. volume
effectb
(L/min)
5.5
12
•'.'•: ,5-5 ,
12
9.5
37
9.5
37
Liq. temp.
effect0
(L/min)
6.7
6.7
0.50
0.50
27
27
0
0
Average = -5.7 16 8.6
"Agitation speed effect from fast to slow.
bLiquid volume effect from high to low.
TLiquid temperature effect from cold to hot.
Table 6-22. Ethylbenzene KLA values for washing machine wash/rinse cycle—Factorial #1
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
?
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Clothes
?
no
no
no
no
no
no
Yes
Yes
Yes
Yes
KLA
(L/min)
10
8.1
17
2.2
2.6
4.3
1.1
4.0
0.93
2.2
Clothes
effect3
(L/min)
8.0
13
1.5
2.1
8.0
13
1.5
2.1
Detergent
effectb
(L/min)
6.7
13
6.7
13
0.1-7
1.8
0.17
1.8
Liq. temp.
effect0
(L/min)
7.9
7.9
1.9
1.9
2.9
2.9
1.3
1.3
Average = 6.1 5.3 3.5
"Clothes effect from full to none.
""Detergent effect from 40 grams to none.
TLiquid temperature effect from cold to hot.
6-34
-------
Table 6-23. Ethylbenzene KLA values for washing machine wash/rinse cycle—Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Gold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
KLA
(L/min)
10
8.1
17
3.0
3.2
2.9
12
38
1.5
1.7
Average =
Agitation
effect"
(L/min)
.y
-21 .
1/r
.O
1.2
-2.9
-21
1.6
1.2
-5.3
Liq. volume
effectb
(L/min)
.0
14
6f\
.0
14
11
36
11
36
17
Liq. temp.
effect0
(L/min)
.9
7.9
.20
-0.20
26
26
0.20
0.20
8.5
"Agitation speed effect from fast to slow.
"Liquid volume effect from high to low.
"Liquid temperature effect from cold to hot.
Table 6-24. Cyclohexane KLA values for washing machine wash/rinse cycle—Factorial #1
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Liquid
Temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Detergent
9
No
No
No
Yes
Yes
Yes
No
No •
Yes
Yes
Clothes
7
No
No
No
No
No
No •
Yes
Yes
Yes
Yes
KLA
(L/min)
24
23
46
9.4
9.2
24
2.9
6.8
3.6
6.0
Clothes
effect"
(L/min)
21
39
5.7.
18
21
39
5.7
21
Detergent
effect6
(L/min)
15
22
15
22
-0.7
0.8
-0.7
0.8
Liq. temp.
effect0
(L/min)
22
22
15
15
3.9
3.9
2.4
2.4
Average = 21 9.2 11
"Clothes effect from full to none.
"Detergent effect from 40 grams to none.
°Liquid temperature effect from cold to hot.
6-35
-------
Table 6-25. Cyclohexane KLA values for washing machine wash/rinse cycle—Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Liquid
temp.
Cold
Cold
Hot
Cold
Cold
Hot
Cold
Hot
Cold
Hot
Liquid
volume
Low
Low
Low
High
High
High
Low
Low
High
High
Agitation
speed
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
KLA
(L/min)
24
23
46
3.4
5.2
4.5
52
94
2.9
5.9
Agitation
effect"
(L/min)
-28
-48
1.4
-1.4
-28
-48
1.4
-1.4
Liq. volume
effectb
(L/min)
20
42.
20
42
49
88
49
88
Liq. temp.
effect0
(L/min)
22
22
0.2
0.2
42
42
3.0
3.0
Average = -19 50 17
"Agitation speed effect from fast to slow.
bLiquid volume effect from high to low.
'Liquid temperature effect from cold to hot.
The second largest main effect on acetone KLA values was use of detergent, with a value of
0.087 L/minute. Regrouping experiments according to water temperature and detergent use
resulted in the following average KLA values: 0.039 L/minute for cold water and no detergent,
0.20 L/minute for hot water and no detergent, 0.0090 L/minute for cold water and detergent, and
0.052 L/minute for hot water and detergent. As shown by these average values, operating
conditions influence the appropriate selection of KLA.
The highest main variable effect for the second factorial group was 0.18 L/minute, again for
liquid temperature. Grouping acetone results according to this main effect resulted in an average
value of KLA of 0.035 L/minute for cold water experiments and 0.21 L/minute for hot water
experiments. The dominance of liquid temperature effects on acetone KLA values for both
factorial groups illustrates the importance of this factor.
Values of KLA for replicate experiments were also compared. For experiments 1(A) and
1(A) replicate, the relative difference in KLA values was 97%. For Experiments 3 and 3
replicate, the relative difference in values of KLA was 27%. Finally for Experiments C and C
replicate, the relative difference in values of KLA was 18%. For wash/rinse cycles, acetone had
relatively low values of KLA, which resulted in larger relative differences. For example,
6-36
-------
acetone's KLA values for Experiments 1 (A) and 1 (A) replicate differed by only 0.0445
L/minute, wMch resulted in a 97% relative difference.
Measured and predicted liquid-phase and gas-phase concentrations for Experiment 6 are
presented in Figure 6-6, and are representative of other experiments. The operating conditions
used in Experiment 6 were hot water, low water volume, slow agitation speed, clothes, and rinse
cycle (no detergent present). As described in Section 6.2.3.3, values of KLA for acetone were
determined by fitting the gas-phase predicted concentrations to the measured gas-phase data for
points collected after 180 seconds into the experiment. As shown in Figure 6-6, the
experimental time of 180 seconds was set to time 0, and the remaining data were also shifted by
180 seconds. The best-fit value of K^A for acetone for this experiment was 0.099 L/minute.
The corresponding hot water wash cycle KLA was 0.082 L/minute. When cold water was used,
the associated wash and rinse cycle values of KLA were 0.0075 L/minute and 0.024 L/minute,
Respectively. • - . - .
At 180 seconds into each experiment (zero in Figure 6-6), the liquid-phase concentration of
acetone was observed to slowly decrease because of the relatively low value of KLA. Figure 6-7
Iff _
35
in t
30
M 25 -
Concentration I
H* h* >>)
O C/l O
w
5 -
Ox
7
i
l_ • ' - : ' _Q_ '_• '
° Measured Liquid Values
Liquid Model Prediction
X Measured Gas Values
1 Gas Model Prediction
' \ - IV 1 V 1
s. i • i /s, i s{ \
) 100 200 300 400 500
Time (seconds)
61
90
Figure 6-6. Acetone concentrations from experiment 6,
6-37
-------
shows a magnification of the y-axis.in Figure 6-6 to illustrate the general decrease in the gas-
phase concentration of acetone during the experiment. The high ventilation rate for washing
machines precluded an approach to chemical equilibrium for all tracers, including acetone.
Values ofKLAfor ethyl acetate ranged from 0.039 to 0.82 L/minute for factorial group #2, as
shown in Table 6-19. Again, the detergent effect on ethyl acetate's elution from the GC
negated the use of factorial #1 experiments in the data analysis. The highest value of KLA was
for the experimental conditions of hot water, low water volume, no clothes or detergent present,
and fast agitation. As with acetone, the largest main effect was liquid temperature with a value
of 0.38 L/minute. The average cold water value of K^A for ethyl acetate was 0.077 L/minute,
and the average hot water value was 0.45 L/minute. Based on the factorial analysis, values of
KLA for ethyl acetate tended to increase with increasing temperature and agitation speed, and
decrease with higher water volumes.
Replicate values of KLA for ethyl acetate had a relative difference of 69% for Experiments
A and A replicate, and 30% for Experiments C and C replicate. Again, the relatively small
values of KLA led to larger relative differences than generally observed for toluene,
ethylbenzene, and cyclohexane.
n n=.
Concentration (mg/L)
s p p p p c
s o o a o i
S M bJ W JX I
i t VX - t
/
X
X Measured Gas Values
Gas Model Prediction
X
0 100 200 300 400 500
Time (seconds)
600
Figure 6-7. Amplification of Figure 6-6 for acetone gas-phase data.
6-38
-------
As shown in Tables 6-20 and 6-21, values ofKLAfor toluene ranged from 0.58 to 38
L/minute, a range covering two orders of magnitude. Similar to the acetone and ethyl acetate
experiments, the operating conditions of hot water, low water volume, no detergent or clothes,
and fast agitation resulted in the highest value of KLA. Unlike acetone and ethyl acetate, the
largest main effect for toluene associated with factorial #1 was presence of clothes, with a value
of 5.3 L/minute. Detergent's main effect was similar to the clothes effect at 5.0 L/minute. As
with stripping efficiency, both of these factors appeared to decrease values of KLA for toluene.
Grouping values of KLA for toluene according to use of detergent and clothes in the
experiment resulted in the following averages: 11 L/minute for no clothes or detergent present,
2.5 L/minute for only detergent present, 2.4 L/minute for only clothes present, and 1.3 L/minute
for both clothes and detergent present. Individually, detergent and clothes had a similar effect on
values of KLA for toluene. These effects appeared to be compounded when both were present in
the machine to lower KLA.
For factorial #2, the liquid volume main effect (16 L/minute) was approximately three times
as high as the main effect associated with agitation speed (-5.7 L/minute), and approximately
two times as high as the main effect associated with liquid temperature (8.6 L/minute). The
average value of KLA was 2.4 L/minute for a high water volume as opposed 16 L/minute for a
low liquid volume.
Values of KLA for replicate experiments were also compared. For experiments 1(A) and
1(A) replicate, the relative difference in values of KLA was 28%. For Experiments 3 and 3
replicate, the relative difference in values of KLA was 50%. Finally, for Experiments C and C
replicate, the relative difference in values of KLA was 7.1%.
Toluene results for Experiment 6 are presented in Figure 6-8. Toluene KLA values were
determined by fitting the predicted liquid concentrations to the measured liquid-phase
concentrations. The best-fit KLA value for this experiment was 3.9 L/minute. The y-axis in
Figure 6-8 is magnified to illustrate the general decrease in toluene gas-phase concentration after
6-39
-------
c
<
,-* ^ "
£
a
d J
o
C3
J-i
S 2-
u
e
o
U
1 -
~*"*^-.
"~~~~._^
o Measured Liquid Value
- - — Liquid Model Predictio
X Measured Gas Values
o ~" ~-
~ - ^.
~~ — • ^
-
;
s
i
0 100 200 300 400 500 600
Time (seconds)
Figure 6-8. Toluene concentrations for Experiment 6.
the initial 180 seconds of the experiment. Like other chemicals, the general shape of the gas-
phase curve for the entire experiment included an increase in gas-phase concentration to a peak,
followed by a decrease in gas-phase concentration as shown in Figure 6-9.
X Measured Gas Values
Gas Model Prediction
1
100 200 300 400
Time (seconds)
500
600
Figure 6-9. Magnification of Figure 6-8 to illustrate toluene's gas-phase
concentration over time.
6-40
-------
Values ofKLAfor ethylbenzene ranged from 0.93 to 38 L/minute for both factorial groups
(see Tables 6-22 and 6-23). Again, this range is similar in magnitude to that of toluene, despite
some difference in Henry's law constant at higher temperatures. Ethylbenzene also had main
effects similar to those calculated for toluene. Based on these main effects, the average
ethylbenzene KLA value for experiments using no detergent or clothes was 12 L/minute. When
detergent or clothes were added to the machine, the average values of KLA were 3.0 L/minute
and 2.6 L/minute, respectively. Finally, when both clothes and detergent were added to the
machine together, the average value of KLA was 1.6 L/minute. .
Values of KLA for ethylbenzene in the second factorial group were most dependent on
liquid volume. An average KLA for ethylbenzene during high water volume experiments was
2.5 L/minute, and an average low water volume KL A for ethylbenzene was 17 L/minute, a
difference of a factor of 7.
Comparing results for replicate experiments yielded the following relative differences in
values of KLA for ethylbenzene: 21% for Experiments 1(A) and 1(A) replicate, 17% for
Experiments 3 and 3 replicate, and 6.5% for Experiments C and C replicate.
Ethylbenzene data for Experiment 6 are plotted in Figure 6-10. Liquid-phase and gas-phase
curves have the same shape as those for toluene. The ethylbenzene KLA value for this plot was
4.0 L/minute.
Finally, values of KLA for cyclohexane ranged from 2.9 L/minute to 94 L/minute for both
factorial groups listed in Tables 6-24 and 6-25. Cyclohexane has a relatively high Henry's law
constant compared with other tracers, which led to consistently higher values of KLA. For these
experiments, there appeared to be significant gas-phase resistance to mass transfer evident by the
wide range of results between tracers.
The greatest main effect for cyclohexane based on factorial #1 was the presence of clothes.
The main effect value of 21 L/minute for clothes was twice as high as the main effect associated
with detergent or water temperature. For the second factorial, the largest main effect was again
6-41
-------
3.5-1
{
3.0"
Ml *> *! -
&
§ 2.0-
B
2
a 1.5-
o
3
M 1-0-
u
0.5-
\
o.o-l
- ~— _
~~--.^
° Measured Liquid Values
Liquid Model Prediction
X Measured Gas Values
f^ -*ir i » -Th T A*
~" --.____
o ~~~--.__^
-~^^ ^
~~ ~~ ~- . — ___
~~ ' — — ^.
/
^""- XX . XX
n
0 100 200 300 400 500 600
Time (seconds)
Figure 6-10. Ethylbenzene concentrations for Experiment 6.
water volume. Average values of KLA for low water volume and high water volume were 48
L/minute and 4.4 L/minute, respectively, a difference of a factor of 1 0 between averages.
Comparing results for replicate experiments yielded the following relative differences in
values of KLA for cyclohexane: 4.3% for Experiments 1(A) and 1(A) replicate, 2.2% for
Experiments 3 and 3 replicate, and 42% for Experiments C and C replicate.
Cyclohexane experimental data are plotted in Figure 6-11 for Experiment 6. The liquid-
phase curve shown in Figure 6-11 has a steeper slope than observed for toluene and
ethylbenzene. The value of KLA for this experiment was 6.8 L/minute for cyclohexane. The
gas-phase curve followed the same shape as for the other tracers.
6.2.4.4. Liquid- and Gas-Phase Mass Transfer Coefficients
To apply the reported values of KLA to other chemicals, it is necessary to separate KLA into
liquid- and gas-phase values ( i.e., kjA, and kgA), and to determine kg/k] for each experiment.
For this system, values of kg/k, should not vary significantly between volatile chemicals. Values
of k]A and kgA for each chemical tracer are listed in Tables 6-26 and 6-27. A single value of
6-42
-------
1 n . ..--.... ....'.
Concentration (mg/L)
» p p p p ;
s t>> 4x b\ bo c
**/ i i i i f\
X\^
*•*
o "^~ • — . ^
o Measured Liquid Value
Liquid Model Predictio
X Measured Gas Values
""— *-— .
"•" I ' 1 r^ 1 IT* 1
0 100 200 300 400 500
Time (seconds)
i
1
^=
6(
)0
Figure 6-11. Cyclohexane concentrations for Experiment 6.
kg/k, is presented based on all chemical tracer experimental values of KLA and physicochemical
properties, as described in Section 3.6.3.
The impact of operating conditions on k, A and kgA was investigated for both factorial
groups as outlined in Section 3.7. For factorial group #1, the most significant factor affecting
k,A for all chemicals was presence of clothes. This result is similar to that of KLA, where the
most significant factor was presence of clothes for all chemicals except acetone (most affected
by temperature). The most significant factor affecting kgA for all chemicals was use of
detergent. For factorial group #2, the most significant factor affecting kjA and kgA for all
chemicals was water volume. These results for toluene, ethylbenzene, and cyclohexane are
similar to those for KLA. The values of KLA for acetone and ethyl acetate were more
significantly affected by temperature. As seen with the shower factorial analysis, there was
typically less dependence on temperature for kgA than for ^A.
As shown in Tables 6-26 and 6-27, the ratio of kg/k, for washing machine wash/rinse cycles
ranged from 0.13 to 8.6, with an average value of 2.2 for factorial group #1 and 2.4 for factorial
group #2. These are relatively low values of kg/k, and are similar in magnitude to values
reported by Hsieh et al. (1994) for diffused bubble aeration.
6-43
-------
Liquid- and gas-phase mass transfer coefficients may also be used to determine the relative
importance of liquid- and gas-phase resistances to mass transfer for specific chemicals and
operating conditions. As shown in Equation 2.5, the overall resistance to mass transfer (1/KLA)
may be written as the sum of liquid-phase resistance to mass transfer (l/k]A) and gas-phase
resistance to mass transfer (l/kgA«Hc). These resistances are shown graphically in Figure 6-12
for each chemical (except ethyl acetate) in Experiment 6. The operating conditions for
Experiment 6 included hot water, low water volume, clothes, no detergent, and slow agitation.
As shown in Figure 6-12, resistance to mass transfer is predominantly gas-phase resistance
dominated for acetone. In fact, the y-axis was adjusted for this plot, because acetone's overall
resistance to mass transfer was much higher (9.8 minutes/L) than the other three chemicals.
Although toluene and ethylbenzene had similar overall resistances to mass transfer for this,
experiment, their respective liquid- and gas-phase resistances to mass transfer were distributed
differently. Gas-phase resistance to mass transfer was slightly greater than liquid-phase
resistance for toluene. With a higher Henry's law constant for this experiment, gas-phase
resistance to mass transfer was smaller than liquid-phase resistance for ethylbenzene. Finally,
gas-phase resistance to mass transfer was insignificant for cyclohexane.
Acetone
Toluene Ethylbenzene Cyclohexane
Chemicals
I Liquid-Phase Resistance (min/D^ Gas-Phase Resistance (min/L)
Figure 6-12. Liquid and gas-phase resistances to mass transfer for Experiment 6.
6-44
-------
6.2.4.5. Mass Closure
For washing machine wash/rinse cycle experiments, mass closure for each chemical was
calculated using Equation 3.10 and based on liquid- and gas-phase measurements collected
during the same period in which values of KLA were determined. Mass closure was reported in
terms of the percentage of mass recovered based on initial total mass.
Values of mass closure for acetone ranged from 95% to 104%, with an average value of
99% for all 17 experiments. Percentages representing mass closure for ethyl acetate ranged from
98% to 114%, with an average value of 104% for applicable experiments (factorial #2
experiments). Mass closure values for toluene ranged from 65% to 135%, with an overall
average of 89%. Ethylbenzene mass closure percentages ranged from 49% to 132%, with an
overall average of 83%. Finally, cyclohexane had a mass closure range of 27% to 137%, with an
average overall value of 72%.
As discussed in Section 4.4.4, mass closure for more volatile chemicals (toluene,
ethylbenzene, and cyclohexane) may be affected by differences in liquid-phase calibration
curves based on tracer bag ages. Again, the actual calibration slope does not affect
determination of chemical stripping efficiencies or values of KLA. It does, however, affect
determination of mass closure for each chemical because of the relation between gas- and liquid-
phase mass. As shown in Section 4.4.4 for showers, improving the liquid-phase calibration
curve resulted in as much as a 15% improvement for toluene mass closure values, a 30%
improvement for ethylbenzene values, and 39% improvement for cyclohexane values.
6-45
-------
Table 6-26. Liquid- and gas-phase mass transfer coefficients for washing machine
wash/rinse cycle experiments—Factorial #1
Experiment
#
1
1 replicate
2
3
3 replicate
4
5
6
7
8
Chemical
A
T
EB
C
A
T
EB
C
A
T
EB
C
A
T
EB
C
A
T
EB
C .
A
T
EB
C
A
T
EB
C
A
T
EB
C
A
T
EB
C
A
T
EB
C
k,A
(L/min)
29
27
28
26
24
38
40
26
56
38
31
49
13
9.6
12
11
9.3
15
15
11
32
49
31
34
2.9
1.3
1.6
3.0
7.7
6.5
5.4
6.9
5.6
2.2
3.6
4.0
5.9
3.4
2.9
6.2
k^
(L/min)
57
.. 53
54
50
22
35
36
34
67
46
37
58
10
7.1
9.2
8.4
7.5
12
12
8.8
4.3
6.6
4.2
4.6
24
11
13
25
21
18
15
19
9.0
3.6
5.7
6.4
18
11
8.9
19
Vk,
1.9
0.92
1.2
0.74
0.81
0.13
8.6
2.8
1.6
3.1
6-46
-------
Table 6-27. Liquid-and gas-phase mass transfer coefficients for washing machine
wash/rinse cycle experiments—Factorial #2
Experiment
#
A
A replicate
B
C
C replicate
D
E
F
G
H
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
k,A
(L/min)
40
23
34
34
26
30
23
46
47
27
62
41
41
33
49
6.4
3.4
5.6
6.1
3.6
9.7
4.8
8.8
9.1
5.6
6.0
3.5
4.4
3.4
4.5
99
45
103
110
68
120
103
161
101
105
4.2
2.4
2.5
2.6
3.0
6.3
3.2
2.5
2.2
6.1
kgA
(L/min)
57
32
49
49
37
22
17
34
35
20
67
45
44
36
53
24
12
21
23
13
20
9.5
18
19
12
30
18
23
17
23
49
22
51
55
34
69
59
92
58
61
27
15
16
17
19
18
8.8
7.0
6.3
17
yk.
1.4
0.74
1.1
3.7
2.1
5.1
0.50
0.58
6.4
2.8
6-47
-------
-------
7. BATHTUB EXPERIMENTS
Three primary activities associated with bathtub operation can cause chemicals originating
in tap water to volatilize: (1) when water flows through a tub spout with an open drain (flow-
through), (2) when water fills the tub with the drain closed (fill), and (3) when the tub is filled
with water (surface volatilization). Bathtub experiments were divided into these three groups.
Bathtub flow-through experiments are described in Section 7.1, fill experiments are discussed in
Section 7.2, and surface volatilization experiments are presented in Section 7.3.
7.1. BATHTUB FLOW-THROUGH EXPERIMENTS
7.1.1. Experimental System
The same shower/bathtub unit described in Section 4.1 was used for all bathtub
experiments. For flow-through experiments, the system had the same modifications and sample
locations as the shower system. As shown in Figure 7-1, the only difference was that the
washing machine contents were pumped through the bathtub spout rather than the showerhead.
7.1.2. Experimental Design
Similar to shower experiments, bathtub flow-through experiments were designed to last 8
minutes. Experimental variables were limited to water temperature and liquid flowrate. To test
all combinations of these conditions, four experiments and one replicate experiment were
completed.
7.1.3. Source-Specific Methodology
Bathtub flow-through experiments followed the same experimental methodology as for
shower experiments (see Section 4.3).
7.1.3.1. Sample Schedule
It was expected that flow-through experiments would have less chemical volatilization than
shower experiments. Therefore, 11 gas samples were collected for 1 minute instead of 30
seconds. In order to collect liquid samples at the mid point of gas sample collection times, the
7-1
-------
1.3 cm OD
Teflon tubing
Gas Sample
Port#l
M/
Chamber
Exhaust Vent
Tedlar Shower
Curtain .
Gas Sample
Port #3
Liquid Tracer Reservoir
Temp Probe
Liquid Sample
Port P
Cinder Block Platform
Figure 7-1. Bathtub flow-through experimental system.
liquid sample schedule was adjusted to 0.5,1.5, two samples at 4.25 and 7.5 minutes. Anine
liquid samples were collected for each experiment.
7.1.3.2. Ventilation Rate
Prediction of bathtub flow-through ventilation rates followed the same procedure developed
for showers (see Section 4.3.2).
7.1.3.3. Parameter Estimation
The only difference between the parameter estimation for bathtub flow-through experiments
and that for shower experiments (see Section 4.3.3) was the method to predict values of KLA for
7-2
-------
acetone. For showers, values of KLA for acetone were predicted based on minimizing the
square of the normalized residual between measured and predicted liquid concentrations. For
bathtubs, volatilization of acetone was near the average duplicate liquid sample error (see
Section 3.5.1), such that the value of KLA was determined using gas-phase data. As for showers,
values of KLA for the remaining tracers were based on liquid-phase data.
7.1.4. Bathtub Flow-Through Results
Six flow-through experiments were completed for which chemical stripping efficiencies and
mass transfer coefficients (KLA, kjA, kgA, and kg/kj) were determined. In addition to these
results, the effects of liquid temperature and liquid flowrate on chemical volatilization are
described in this section.
The operating conditions for each flow-through experiment are listed in Table 7-1.
7.1.4.1. Chemical Stripping Efficiencies
Chemical stripping efficiencies (ti) are reported in Table 7-2. Stripping efficiencies for each
chemical were based on liquid-phase measurements.
Acetone stripping efficiencies ranged from 1.7% to 5.3%. The highest value corresponded
to the conditions of high flowrate and warm water. In fact, the two experiments completed with
warm water led to the highest stripping efficiencies for acetone. Grouping stripping efficiencies
according to water temperature and averaging them resulted in a cold water average of 2.9% and
a warm water average of 4.9%.
Table 7-1. Bathtub flow-through operating conditions
Experiment
#
1
1 replicate
2
3
4
4 replicate
Liquid
temperature
(°O
22
23
36
25
36
37
Liquid
flowrate
(L/min)
9.1
9.1
9.1
6.1
6.1
6.1
Gas
flowrate
(L/min)
355
345
359
350
361
365
ACH
(1/hr)
12
12
12
12
12
13
7-3
-------
Table 7-2. Chemical stripping efficiencies (r\) for experimental bathtub flow-through
ex
Experiment^
1
1 rep
2
3
4
4 rep
periments
Liquid
temp.
Cold
Cold
Warm
Cold
Warm
Warm
Liquid
flowrate
High
High
High
Low
Low
Low
Acetone
n (%)
3.8
3.1
5.3
1.7
4.6
4.8
Ethyl acetate
TI(%)
6.1
4.7
11
4.5
14
10
Toluene
n (%)
26
24
38
22
30
38
Ethylbenzene
Tl (%)
27
24
39
22
29
38
Cyclohexane
Tl (%)
28
29
38
22
27
41
Ethyl acetate stripping efficiencies ranged from 4.5% to 14% and followed the same trends
as acetone. A cold water average stripping efficiency for ethyl acetate was 5.1% and a warm
water average stripping efficiency was 12%. The trend of increasing stripping efficiency with
increasing temperature is primarily caused by the resulting increase in Henry's law constant for
each chemical.
The stripping efficiencies for toluene, ethylbenzene, and cyclohexane were of similar
magnitude. The ranges of stripping efficiencies for each chemical were 22% to 38% for toluene,
22% to 39% for ethylbenzene, and 22% to 41% for cyclohexane. The fact that toluene and
ethylbenzene results were similar was not surprising given their similar Henry's law constants.
For the range of temperatures listed in Table 7-1, toluene had Henry's law constants of 0.25
m3i;q/m3gas (Experiment 1) to 0.39 m3liq/m3gas (Experiment 4), and ethylbenzene had Henry's law
constants of 0.28 m3,iq/m3gas (Experiment 1) to 0.60 m3Iiq/m3gas (Experiment 4). The fact that
cyclohexane also had results similar to toluene and ethylbenzene indicated that there was not
significant gas-phase resistance to mass transfer for the more volatile chemicals for this system.
For the temperatures listed hi Table 7-1, cyclohexane had significantly higher Henry's law
constants (6.6 m^/m3^ [Experiment 1] to 11 m3liq/m3gas [Experiment 4]) than did either toluene
or ethylbenzene.
The temperature-dependent average stripping efficiencies for toluene, ethylbenzene, and
cyclohexane were as follows: cold water averages were 24%, 24%, and 26%, respectively; and
warm water averages were 35% for all three chemicals.
7-4
-------
Bathtub flow-through Experiment 1 and Experiment 4 were repeated and reported as
Experiment 1 replicate and Experiment 4 replicate, respectively. The relative difference
between stripping efficiencies determined for Experiment 1 and Experiment 1 replicate for each
chemical was 20% for acetone, 26% for ethyl acetate, 8.0 % for toluene, 12% for ethylbenzene,
and 3.5% for cyclohexane. The relative difference between stripping efficiencies determined for
Experiment 4 and Experiment 4 replicate for each chemical was 4.3% for acetone, 33% for ethyl
acetate, 24% for toluene, 27% for ethylbenzene, and 41% for cyclohexane.
7.1.4.2. KLA Values
Values of KLA for all chemicals are reported in Table 7-3. Values of KLA for acetone were
based on gas-phase data. Values of KLA for ethyl acetate, toluene, ethylbenzene, and
cyclohexane were based on liquid-phase data.
The highest values of KLA for acetone and ethyl acetate were associated with warm water
experiments. Values of KLA for acetone ranged from 0.11 to 0.54 L/minute. The cold water
average value of KLA was 0.15 L/min and the warm water average was 0.48 L/min. Values of
KLA for ethyl acetate ranged from 0.32 to 1.2 L/minute. The temperature dependent averages of
KLA for ethyl acetate were 0.48 L/min for cold water and 1.0 L/min for warm water.
The highest values of KLA for toluene, ethylbenzene, and cyclohexane were for the
conditions of high flowrate and warm water. However, experiments using cold water and a high
flowrate also resulted in higher values of KLA. Because the range of values of KLA was so
Table 7-3. Values of KT A for bathtub flow-through experiments
Experiment
#
1
1 rep
2
3
4
4 rep
Liquid
temp.
Cold
Cold
Warm
Cold
Warm
Warm
Liquid
flowrate
High
High
High
Low
Low
Low
Acetone
KLA
(L/min)
0.11
0.15
0.54
0.18
0.43
0.46
Ethyl acetate
KLA
(L/min)
0.64
0.49
1.2
0.32
1.1
0.79
Toluene
KLA
(L/min)
2.9
2.4
4.5
1.6
2.2
2.9
Ethylbenzene
KLA
(L/min)
2.9
2.4
4.5
1.5
2.1
2.9
Cyclohexane KLA
(L/min)
3.1
2.9
5.1
1.7
1.9
3.2
7-5
-------
narrow for these compounds, an overall average is reported here. Values of KLA for toluene
ranged from 1.6 to 4.5 L/minute, with an overall average of 2.8 L/minute. Values of KLA for
ethylbenzene ranged from 1.5 to 4.5 L/minute, with an overall average of 2.7 L/minute. Finally,
values of KLA for cyclohexane ranged from 1.7 L to 5.1 L/minute, with an overall average of 3.0
L/minute.
Mass transfer data for bathtub flow-through experiments may be presented in the same
format as shower experimental data (see Section 4.4.2). A representative plot is shown hi Figure
7-2 for toluene and Experiment 4 replicate. The operating conditions used in Experiment 4
replicate were warm water and low flowrate. As shown in Figure 7-2, each experimental period
consisted of a liquid sample collected from the tracer reservoir, an outlet liquid sample, and a gas
sample. For each period, the bathtub outlet concentration in both the liquid and gas phases may
be estimated using the mass balance models (Equations 2.28 and 2.30). To determine the best
KLA value for the model, the residuals between the measured and predicted concentrations were
minimized using the method described in Section 3.6.2. For toluene, the residual between
liquid-phase values was minimized, resulting in a value of 2.9 L/minute for Experiment 4
replicate.
o
ra
JS
•*
S
Initial
7 Period
M 6 -
a
ts 4'
£ 3 -
J|2-
3 1 -
° 0-
Intermediate Final
'O
~x.x.
^
/£
^--
0 75 12C
Period
0
— —
-^*^***^-"^ 5^
X
165 .210 255 300
Period
0
_____ —
X
X^v
. _._. Jf-
0.030
0.025 «
- 0.020 |
- 0.015 g
- 0.010 «
- 0.005
n nnn
345 390 435
Time (seconds)
O Measured Reservoir Liquid Values >K Measured Liquid Values
• Model Predicted Liquid Values X Measured Gas Values
- Model Predicted Gas Values
g
•S
c
%
1
O
u
Figure 7-2. Toluene experimental data for Experiment 4 replicate.
7-6
-------
7.1.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients
Values of KLA for each chemical were separated into the components of k,A and kgA using
Equation 2-5 and a value" of kg/k, determined for each specific experiment (see Sections 3.6.3 and
3.6.4 for methodology). Values of k,A and kgA are reported in Table 7-4 for each chemical in
addition to values of kg/k, for each experiment. •
For bathtub flow-through experiments, values of kg/k, ranged from 37 to 96. In general,
values of k,A and kgA for each chemical were similar in magnitude. The ratio of kg/k, was higher
at low fiowrates (average kg/k, = 71) than at high flowrates (average kg/k, = 41).
Table 7-4. Liquid- and gas-phase mass transfer coefficients for bathtub flow-through
experiments
Experiment
#
1
1 replicate
2
3
4
4 replicate
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
k,A (L/min)
2.9
4.5
3.2
3.1
3.1
3.2
3.0
2.6
2.6
2.9
6.0
4.9
4.8
4.7
5.1
2.4
1.3
1.7
1.6
1.7
2.4
2.4
2.2
2.2
1.9
4.2
2.7
3.1
3.0
3.2
kgA (L/min)
108
168
117
117
115
136
126
111
109
126
249
205
198
196
211
159
86
110 .
106
110
227
234
214
207
182
211
135
152
151
161
kg/k,
37
43
42
66
96
50
7-7
-------
Liquid- and gas-phase mass transfer coefficients may also be used to determine the relative
importance of liquid and gas-phase resistances to mass transfer for specific chemicals and
operating conditions. As shown hi Equation 2-5, overall resistance to mass transfer (1/KLA) may
be written as the sum of liquid-phase resistance to mass transfer (1/kjA) and gas-phase resistance
to mass transfer (l/kgA»Hc). These resistances are shown graphically in Figure 7-3 for each
chemical in Experiment 2. As shown in Figure 7-3, overall resistance to mass transfer for
acetone and ethyl acetate is dominated by resistance in the gas phase. On the other hand, overall
resistance to mass transfer for toluene, ethylbenzene, and cyclohexane is insignificant compared
with their respective liquid-phase resistances to mass transfer.
7.1.4.4. Mass Closure
The ranges of mass closure for each chemical were 98% to 102% for acetone, 98% to 105%
for ethyl acetate, 89% to 107% for toluene, 86% to 100% for ethylbenzene, and 82% to 103% for
cyclohexane. All mass closure values are given in the Appendix.
o.o
Acetone Ethyl Acetate Toluene Ethylbenzene Cyclohexane
Chemicals
I Liquid-Phase Resistance (min/I{5 Gas-Phase Resistance (min/I,)
Figure 7-3. Resistances to mass transfer for each chemical in Experiment 2.
7-8
-------
7.2. BATHTUB FILL EXPERIMENTS
7.2.1. Experimental System
As before, the same experimental system constructed for showe^athtub flow-through
experiments was used for bathtub fill experiments. However, the drain was plugged with a
rubber stopper so the bathtub would fill. An additional modification shown in Figure 7-4 was a
different liquid sample port location. Samples were collected from this port by pumping water
from the bathtub using a 102 cm perforated 0.635 cm OD Teflon ™ tube. The perforated
Teflon™ tube was angled in the bathtub pool such that water was drawn through the holes at
different depths, resulting hi a more distributed liquid sample. Tracer reservoir samples were
collected in the same manner as for shower and bathtub flow-through experiments. For bathtub
fill experiments, only gas sample ports #1 and #2 were used. A more evenly distributed gas
concentration was expected for these experiments.
7.2.2. Experimental Design ,
As with flow-through experiments, bathtub fill experimental variables were liquid
temperature and liquid flowrate. -Four experiments were completed with two replicates.
7.2.3. Source-Specific Methodology
No different preexperimental tasks were completed for fill experiments, except for the
introduction of a drain plug.
7.2.3.1. Sample Schedule
Bathtub fill experiments varied in length depending on experimental flowrate. For
experiments completed at 9.1 L/min, the bathtub was filled for 8 minutes. For experiments
completed at 6.1 L/min, the bathtub was filled for 12 minutes. A total of 10 liquid-phase
samples were collected for each fill experiment. Liquid-phase samples were collected every 2
minutes for the high flowrate experiments, and every 3 minutes for the low flowrate
experiments. Tracer reservoir samples were collected within 90 seconds of every bathtub
sample.
7-9
-------
1.3 cm OD
Teflon tubing
Gas Sample
Port #2
Bathtub
Spout
Liquid Temp
,, / Probe
HI/
l\ Plugged
\\j/ Drain _
Gas Sample
Port#l
Chamber Exhaust
— Vent
Tedlar Shower
Curtain
Liquid Sample Port
for Filled Basin
Tracer
Reservoir
Liquid
Rotameter
Cinder Block Platform
Figure 7-4. Bathtub fill experimental system.
An initial gas sample was collected from sample port #1 before starting the experiment.
Once an experiment began, a single gas sample was collected at this port for the duration of the
experiment, followed by a 1-minute sample collected immediately after the experiment. One-
minute gas samples were also collected at sample port #2. These samples were scheduled such
that a bathtub liquid sample was collected at the midpoint of the gas sample time. Six gas-phase
samples were collected for each experiment.
7-10
-------
7.2.3.2. Ventilation Rates
Prediction of bathtub (fill) ventilation rates followed the same procedure developed for
showers (see Section 4.3.2).
7.2.3.3. Parameter Estimation
The same parameter estimation techniques outlined in Section 6.1.3.3 for washing machine
fill cycle experiments applied to bathtub fill experiments. Stripping efficiencies and values of
KLA were based on liquid-phase data for all chemicals. Values of KLA were based on gas-phase
data for acetone and ethyl acetate and were based on liquid-phase data for toluene, ethylbenzene,
and cyclohexane.
7.2.4. Bathtub Fill Results
Five bathtub fill experiments were completed to predict chemical mass emissions. Bathtub
fill results can be combined with bathtub surface volatilization results presented in Section 7.3 to
characterize total mass emissions during a typical bathing event. Ventilation rates, stripping
efficiencies, and mass transfer coefficients (KLA, k,A, kgA, and kg/kj are presented in this
chapter and are based on the experimental methodology presented in Sections 3.0 and 7.2.3. In
addition, the effects of liquid temperature, liquid fill rate, and chemical properties are discussed.
The operating conditions for each mass transfer experiment are listed in Table 7-5.
7.2.4.1. Chemical Stripping Efficiencies
Chemical stripping efficiencies (r|) are reported in Table 7-6. Stripping efficiencies for high
flowrate (9.1 L/min) experiments were based on a fill time of 8 minutes whereas stripping
efficiencies for low flowrate (6.1 L/min) experiments were based on a fill time of 12 minutes.
Liquid-phase concentrations did not change significantly in the bathtub after approximately i
minute of fill time, thereby allowing comparison of stripping efficiencies based on different
experimental times.
7-11
-------
Table 7-5. Bathtub (fill) operating conditions
Experiment
#
1
2
2 replicate
3
4
Liquid
temperature
(°C)
24
35
36
23
35
Fill
time
(min)
8:00
8:00
8:00
12:00
11:23
Liquid
flowrate
(L/min)
9.1
9.1
9.1
6.1
6.1
Liquid
volume
(L)
73
73
73
73
69
Gas
flowrate
(L/min)
373
379
373
370
377
Table 7-6. Chemical stripping efficiencies (r\) for bathtub (fill) experiments
Expt.
#
1
2
2 rep
3
4
Liquid
Temp.
Cold
Warm
Warm
Cold
Warm
Liquid
Flowrat
e
High
High
High
Low
Low
Aceton
e
n (%)
4.9
5.2
2.0
5.8
7.7
Ethyl
Acetate r\
(%)
3.0
5.3
3.1
3.1
7.0
Toluene
1 (%)
31
30
31
.29
30
Ethylbenze
ne r| (%)
33
32
32
31
29
Cyclohexan
e
n (%)
46
47
46
43
46
Stripping efficiencies for acetone ranged from 2.0% to 7.7%, with the highest value for low
flowrate and warm water. The average stripping efficiency for acetone was 5.1%. Stripping
efficiencies for ethyl acetate ranged from 3.0% to 7.0%, with an overall average of 4.3%.
There was little deviation between stripping efficiencies for toluene, ethylbenzene, and
cyclohexane. Average stripping efficiencies were 30% for toluene, 31% for ethylbenzene, and
46% for cyclohexane. As expected, toluene and ethylbenzene had similar results. Cyclohexane
stripping efficiencies were somewhat higher for this set of bathtub experiments, indicating more
influence of gas-phase resistance to mass transfer for more volatile chemicals, possibly from
formation of bubbles in the underlying pool as the bathtub filled.
Experiment 2 was repeated. The relative difference between stripping efficiencies
determined for Experiments 2 and 2 replicate were 88% for acetone, 52% for ethyl acetate, 3.3%
for toluene, 0% for ethylbenzene, and 2.2% for cyclohexane.
7-12
-------
7.2.4.2. KLA Values
Values of KLA for all chemicals and operating conditions are reported in Table 7-7, and
were based on the same fill times discussed for stripping efficiencies. In general, for all
chemical tracers, values of KLA were higher at higher liquid flowrates. The average values of
KLA at a liquid flowrate of 6.1 L/min were 0.39 L/min for acetone, 0.86 L/min for ethyl acetate,
2.7 L/min for toluene, 2.6 L/min for ethylbenzene, and 4.9 L/min for cyclohexane. At a liquid
flowrate of 9.1 L/min, the average values of KLA were 0.54 L/min for acetone, 1.3 L/min for
ethyl acetate, 4.4 L/min for toluene, 4.4 L/min for ethylbenzene, and 8.5 L/min for cyclohexane.
7.2.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients
Values of KLA for each chemical were separated into the components of k,A and kgA using
Equation 2.5 and a value of kg/k, determined for each specific experiment. These values are
reported in Table 7-8. For the bathtub filling events, values of kg/k, ranged from 27 to 77, with
an average value of 51. Bathtub water during a filling event is characterized by splashing at the
surface and entrainment of air creating visible bubbles in the pool. As such, the average kg/k,
value for high flowrate experiments was somewhat lower (45) than the average value associated
with low flowrate experiments (62). These values were similar in magnitude to values of kg/k,
predicted for surface aerator systems (38 to 110) (Hsieh et al., 1991).
7.2.4.4. Mass Closure
For bathtub fill experiments, the percentage of mass recovered was based on Equation 3.11
applied for the entire time of fill. The range of mass closure for each chemical was 96% to
101% for acetone, 103% to 108% for ethyl acetate, 89% to 106% for toluene, 87% to 96% for
ethylbenzene, and 68% to 87% for cyclohexane.
Table 7-7. Values of KLA for bathtub (fill) experiments
Experiment
#
1
2
2 rep
3
4
Liquid
temp.
Cold
Warm
Warm
Cold
Warm
Liquid
flowrate
High
High
High
Low
Low
Acetone
KLA
(L/min)
0.45
0.53
0.64
0.39
0.39
Ethyl acetate
KLA
(L/min)
1.0
1.4
1.5
0.71
1.0
Toluene
KLA
(L/min)
4.1
5.3
3.7
2.6
2.7
Ethylbenzene
KLA
(L/min)
4.4
5.9
3.8
2.7
2.5
Cyclohexane
KLA
(L/min)
7.1
11
7.4
4.4
5.4
7-13
-------
Table 7-8. Liquid- and gas-phase mass transfer coefficients for bathtub (fill) experiments
Experiment
#
1
2
2 replicate
3
4
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
k,A
(L/min)
7.1
4.9
4.4
4.6
7.1
9.3
8.3
5.8
6.3
11
5.9
5.3
3.8
4.0
7.4
4.7
2.7
2.7
2.8
4.4
4.2
3.8
2.8
2.6
5.4
(L/min)
395
274
244
257
396
253
228
159
172
311
303
269
193
202
376
365
208
208
220
344
191
175
129
121
245
Vk,
56
27
51
•77
46
7.3. SURFACE VOLATILIZATION EXPERIMENTS
7.3.1. Experimental System
The same experimental system presented in Section 7.2.1 was used for surface volatilization
experiments. The only addition to the system was a simulated person. A model of a person was
designed using empty 3 L Tedlar™ bags. Four bags were connected with string to create joints
between each bag. Additional strings were attached to the "head" bag and "foot" bag that
allowed them to be moved from outside the system. Moving the strings resulted in bag motions
that created waves and moderate splashing in the bathtub pool.
7-14
-------
7.3.2. Experimental Design
Four combinations of conditions were studied for surface volatilization experiments.
Experimental variables were liquid temperature and the presence of a person.
7.3.3. Source-Specific Methodology
Surface volatilization experiments followed bathtub fill experiments, so the bathtub
contained a well-mixed solution of chemical tracers. The following tasks were completed before
starting the surface volatilization portion of an experiment:
• For appropriate experiments, the Tedlar™ person was placed in the bathtub pool
• An initial liquid sample was collected
• An initial gas sample was collected from port #1.
7.3.3.1. Sample Schedule
Based on typical bathing times, surface volatilization experiments lasted 20 minutes.
Liquid samples were collected from the bathtub every 4 minutes. Eight liquid-phase samples
were collected for each experiment. Gas samples were collected from port #2 for 2 minutes
each, starting at 3, 7, 11, and 15. minutes. A single gas sample was collected for the 20-minute
experiment from port #1. A final gas sample was also collected from this port. Six gas samples
were collected for each experiment.
7.3.3.2. Ventilation Rates
Prediction of ventilation rates associated with bathtub surface volatilization experiments
followed the same procedure developed for showers (see Section 4.3.2).
7.3.3.3. Parameter Estimation
Chemical volatilization rates for bathtub experiments with standing water (no additional
motion) were nearly zero. Therefore, for surface volatilization experiments, chemical stripping
efficiencies were calculated using the following equation, which included gas-phase data:
7-15
-------
v,cun
where
Qg = system ventilation rate (L3/T)
Cg = integrated gas sample average concentration (M/L3)
At = time of experiment (T)
Vg = headspace volume (L3)
V, = bathtub fill volume (L3)
Cg cnd = final gas-phase concentration in headspace (M/L3)
C&in = initial gas-phase concentration in headspace (M/L3)
CI>in = average liquid-phase concentration in tracer reservoir (M/L3).
(7-1)
Bathtub surface volatilization experiments with a simulated person were characterized by
significantly higher chemical stripping efficiencies than were quiescent bathtub experiments,
such that liquid-phase values were used.
Because of the limited chemical volatilization for quiescent conditions (no "person"
present), values of KLA were not determined for this source. Values of KLA for bathtub
experiments with a simulated person were determined using the methodology outlined in Section
3.6.2. Gas-phase data were used to find the best-fit value of KLA for acetone and ethyl acetate,
and liquid-phase data were used to find the best-fit value of KLA for toluene, ethylbenzene, and
cyclohexane.
7.3.4. Bathtub Surface Volatilization Results
Six bathtub surface volatilization mass transfer experiments were completed. Surface
volatilization results may be combined with bathtub fill results presented in Section 7.2.4 to
characterize total mass emissions during typical bathing events. Based on the experimental
methodology presented in Sections 3.0 and 7.3.3, the ventilation rates, overall chemical stripping
efficiencies, and mass transfer coefficients (KLA, k,A, kgA, and kg/k,) are presented in this
7-16
-------
chapter. In addition, effects of liquid temperature, presence of a person, and chemical properties
on each response are discussed.
Operating conditions for each mass transfer experiment are given in Table 7-9.
7.3.4.1. Chemical Stripping Efficiencies
Stripping efficiencies for each chemical are reported in Table 7-10. As mentioned in
Section 7.3.3.3, chemical stripping efficiencies for experiments not using a simulated person
(Experiments 1, 1 replicate, and 2) were based on gas-phase data. Values reported in Table 7-10
for these experiments ranged from 0.57% to 15% for all chemicals.
There were significant reductions in liquid-phase chemical concentrations for Experiments 3
through 4 replicate, allowing stripping efficiencies to be determined based on differences in
liquid concentration. Stripping efficiencies ranged from 1.6% to 7.3% for acetone, 3.4% to 9.8%
for ethyl acetate, 27% to 32% for toluene, 26% to 32% for ethylbenzene, and 30% to 41% for
cyclohexane. The degree of splashing associated with surface volatilization experiments with a
simulated person could not be reasonably quantified and was not consistent. As a result, it is
difficult to report more than just stripping efficiency values for these experiments; that is, neither
a trend analysis nor a determination of relative difference in values was completed.
Table 7-9. Bathtub surface volatilization operating conditions
Experiment
#
1
1 replicate
2
3
4
4 replicate
Liquid
temperature
(°C)
23
24
34
24
33
35
Person
present?
No
No
No
Yes
Yes
Yes
Liquid
volume
(L)
73
73
69
73
73
73
Gas
flowrate
(L/min)
370
377
377
373
379
373
7-17
-------
Table 7-10. Chemical stripping efficiencies (n) for bathtub surface volatilization
experiments
Experiment.
#
1
1 rep
2
3
4
4 rep
Liquid
temp.
Cold
Cold
Warm
Cold
Warm
Warm
Person
present
?
No
No
No
Yes
Yes
Yes
Acetone
T, (%)
0.57
2.5
2.7
1.6
4.5
7.3
Ethyl
acetate r\
(%)
1.6
5.9
6.4
3.4
9.8
8.9
Toluene
TI(%)
7.9
13
14
32
27
30
Ethylbenzene
T! (%>
5.1
7.6
8.3
32
26
29
Cyclohexane
1 (%)
4.7
13
15
39
41
30
7.3.4.2 KLA Values
Values of KLA were not determined for experiments using still water (no person present)
given the near zero rate of chemical volatilization. Values of KLA were determined, however,
for Experiments 3, 4, and 4 replicate and are reported in Table 7-11.
Values of KLA for each chemical were 0.11 to 0.25 L/minute for acetone, 0.24 to 0.49
L/minute for ethyl acetate, 1.2 L/minute for toluene, 1.1 to 1.2 L/minute for ethylbenzene, and
1.2 to 2.0 L/minute for cyclohexane. Again, because of the inconsistent nature of these
experiments, values of KLA are merely indicators of the order of magnitude of chemical
volatilization during bathing events.
A representative experimental plot for surface volatilization experiments with a simulated
person is presented hi Figure 7-5. The plot shows toluene data. Experimental conditions were
warm water and presence of the simulated person. The resulting value of KLA for toluene for
this example was 1.2 L/min.
Table 7-11. Values of KT A for bathtub surface volatilization experiments
Experiment
#
3
4
4 rep
Liquid
temp.
Cold
Warm
Warm
Person
present?
Yes
Yes
Yes •
Acetone
KLA
(L/min)
0.11
0.25
0.23
Ethyl acetate
KLA
(L/min)
0.24
0.49
0.40
Toluene
KLA
(L/min)
1.2
1.2
1.2
Ethylbenzene
KLA
(L/min)
1.2
1.1
1.1
Cyclohexane
KLA
(L/min)
1.4
2.0
1.2
7-18
-------
jf 4 •"
¥ 3-
ta
£ 2
a. 1
«.
-X-
200 400 600
Time (sec)
800
1000
1200
o Measured Liquid Values Liquid Model Prediction
X Measured Gas Values Gas Model Prediction
Figure 7-5. Toluene experimental data for Experiment 4 replicate.
7.3.4.3. Liquid- and Gas-Phase Mass Transfer Coefficients
Values of KLA for each chemical were separated into the components of kjA and kgA using
Equation 2.5 and a value of kg/k] determined for each specific experiment (see Sections 3.6.3 and
3.6.4 for methodology). Values of kjA and kgA are reported in Table 7-12 for each chemical in
addition to values of kg/k, for Experiments 3, 4, and 4 replicate.
For bathtub surface volatilization experiments with a simulated person present, values of
ranged from 54 to 78. Despite the randomness associated with these experiments, values of
were relatively similar in magnitude.
7.3.4.4. Mass Closure
For bathtub surface volatilization experiments, the percentage of mass recovered was based
on Equation 3.10. The range of mass closure for surface volatilization experiments with no
person present was 99% to 104% for acetone, 100% to 105% for ethyl acetate, 96% to 110% for
toluene, 86% to 100% for ethylbenzene, and 93% to 117% for cyclohexane. For surface
volatilization experiments involving a simulated person, the range of mass closure was 98% to
103% for acetone, 104% to 109% for ethyl acetate, 90% to 100% for toluene, 83% to 92% for
ethylbenzene, and 80% to 91% for cyclohexane.
7-19
-------
Table 7-12. Liquid- and gas-phase mass transfer coefficients for bathtub surface
volatilization experiments
Experiment
#
3
4
4 replicate
Chemical
A
EA
T
EB
C
A
EA
T
EB
C
A
EA
T
EB
C
k,A
(L/min)
1.8
1.2
1.3
1.3
1.4
1.8
1.4
1.2
1.1
2.0
1.6
1.1
1.2
1.1
1.2
(L/min)
97
63
69
68
76
143
107
97
88
156
122
84
97
88
94
"A
54
78
78
7-20
-------
8. MODEL APPLICATIONS
All experimental results reported herein as well as previously reported research results have
been compiled into a database that is provided in the Appendix. Each entry in the database
summarizes a particular experiment, including operating conditions, chemical stripping
efficiencies, and, where applicable, estimated values of KL A, kjA, kgA, and percent mass
recovery. The database could serve as a tool for a user to find the most appropriate modeling
parameters for a specific contamination event. At this time, the database includes 164 shower
results (including 50 from this study), 44 dishwasher results (all from this study), 128 washing
machine results (including 114 from this study), 85 bathtub results (all from this study), and 33
kitchen sink results. Using the available information, it is now possible to estimate chemical
emissions .from tap water sources for numerous scenarios, without having to assume 100%
volatilization for all chemicals.
Based on experimental results, values of KLA and, where appropriate, headspace ventilation
rates can be used in conjunction with associated source mass balance models to determine
chemical emissions during a specific source event. In this chapter, an example event for each of
the four sources discussed herein is presented. The methodology for predicting emissions for
other chemicals of interest is provided. For each source, toluene was used as the surrogate
compound (chemical j in Equation 2-15). Dibromochlorornethane (DBCM), a common
disinfection by-product, and methyl ethyl ketone (MEK), a common solvent, were used as the
chemicals of interest (chemical i in Equation 2-15). A comparison of these three chemicals is
provided in Table 8-1. For all cases, chemicals were assumed to be present in the water supply
at a concentration of 10 |_ig/L.
8.1. SHOWER MODEL APPLICATION
Mass balance Equations 2-28 and 2-30 may be used to predict chemical liquid- and gas-
phase concentrations during a shower event of any duration. The associated mass emissions may
be estimated during a shower event by applying the predicted liquid-phase concentrations to
Equation 2-32. For this example, a shower duration of 10 minutes was chosen. Other operating
8-1
-------
Table 8-1. Comparison of the three chemicals used in model applications
Chemical
Toluene
Dibromochloromethane
Methyl Ethyl Ketone
Hc @ 25°C
(myin^)"
0.27
0.048
0.0060
D,@24°C
(cm2/sec)b
9.1 x IQ-6
l.OxlQ-5
9.8 x lO'6
Dg @ 24°C
(cm2/sec)b
0.085
0.086
0.097
'From Ashworth et al., 1988.
bFrom Tucker and Nelken, 1990.
conditions for this example, based on experimental operating conditions, were a water
temperature of 35°C, a liquid flowrate of 9.1 L/minute, and a ventilation rate of 379 L/minute
(resulting in an air exchange rate of 13/hour). A coarse water spray was assumed. The value of
KLA for toluene (used in Equations 2-28 and 2-30 to predict toluene liquid- and gas-phase
concentrations, respectively, associated with these operating conditions) was assumed to be 12
L/minute. This value is the average KLA determined for shower Experiments 5, 6, and 6
replicate (see Section 4.4.2). It should be noted that several values of KLA based on different
shower operating conditions are available in the experimental database (Appendix).
Based on an inlet liquid-phase concentration of 0.010 mg/L and an initial gas-phase
concentration of 0 mg/L, the predicted mass emission rate for toluene is presented in Figure 8-1.
e
1,0.06-
f 0-05 =
H 0.04-
| 0-03 -
,| 0.02-
| 0.01 -
^4>
i
-1 c I 1 1
0 2 46 8 10
Time (minutes)
Toluene H~ Dibromochloromethane a Methyl Ethyl Ketone
Figure 8-1. Mass emission rates for three chemicals for example shower event.
8-2
-------
The total mass of emitted toluene was calculated by integrating under the mass emission
rate curve shown in Figure 8-1. For this example, the total emitted mass of toluene was 650 u.g.
The total mass that entered the system was 910 ug. Thus, the overall stripping efficiency for
toluene during the 10-minute shower event was 71%. The peak gas-phase concentration within
the shower stall occurred at 10 minutes and was approximately 150 u.g/m3.
The mass emission rates for DBCM and MEK, two chemicals not used in this study, are
also shown in Figure 8-1. For toluene and DBCM, the mass emission rate slowly decreased with
time as each chemical accumulated within the shower stall. This effect was more dramatic for
MEK, the chemical with the lower Henry's law constant. The overall stripping efficiencies for
DBCM and MEK were 66% and 13%, respectively.
The procedure for predicting mass emissions for any chemical of interest based on the
results of this study is illustrated by means of a step-by-step method for one chemical of interest,
MEK. The shower conditions described earlier for toluene also apply for this example.
Step 1: Choose an experimental tracer to be the surrogate compound with an associated
value of KLA.
For this example, toluene was chosen as the surrogate compound (chemical j). As
shown earlier, the value of KLA for toluene and associated operating conditions was 12
L/minute.
Step 2: Choose appropriate experimentally determined kg/k, value for source operating
conditions.
The value of kg/k, for any shower event was estimated to be 160 (see Section 4 A3).
Step 3: Estimate *P, for surrogate compound (chemical j) and chemical of interest
(chemical i).
For toluene and MEK, the value of T, was calculated using Equation 2-12 as:
8-3
-------
D
= 1.1
I, toluene
Step 4: Estimate *Fg for surrogate compound (chemical j) and chemical of interest
(chemical i).
For toluene and MEK, the value of Tg was calculated using Equation 2-13 as:
„, I PfeMEK I3 _ -I -I
8 D • A.l
V. g, toluene J
Step 5: Estimate Tm for surrogate compound (chemical j) and chemical of interest
(chemical i).
^¥m was calculated using Equation 2-15 with values from Steps 2 through 4 and Henry's
law constants for each chemical listed in Table 8-1. The values of Henry's law
constant for each chemical were adjusted for a temperature of 35°C using correlations
developed by Ashworth et al. (1988).
_ \ir \
- M
1 +
=(u).(i.i). |«^.l. | 1/^0.0.37 1
V / V ^ \ 0.37 J [1.1 +1.1 • 0.0033• 160J
8-4
-------
Step 6: Calculate KLA for chemical of interest.
The value of KLA for MEK may be estimated using:
= x*'m-KLAtoluene = 0.38 «12 L/minute = 4.5 L/minute.
Step 7: Predict liquid- and gas-phase concentrations as a function of time.
Applying a value of KLA of 4.5 L/minute to Equations 2-28 and 2-30 enables
prediction of liquid- and gas-phase concentrations, respectively, of MEK. At 10
minutes, the gas-phase concentration in the shower stall is the following:
C&10min=|+(c&0-|W-Dt)
2.0x10 5mg/(mm»L)
C = - •»«•»/• - '
g,iOmin 0.83 /mm
2.0xlO~5mg/(min«L)
noof • -
0.83 /mm
- . ,
exp(-0.83/min. lOmin) = 2.4xKr5mg/L
6
where
B = -
V
B =
9.1L/min. 0.01 mg/L »| 1 -expf- 4'5 L'mm] I + 379 L/min • Omg/L
V 9.1 L/min. '
'1745 L
= 2.0 x 10"3 mg / (min • L)
8-5
-------
D = -
Q,
V
9-lJL/min
0.0033
_
4.5L/minY\ __A T , . I
—-——-+379 L / mm
9.lL/rmnJJ ' )
= 0.83mm"1
1745 L
The resulting liquid-phase concentration at the shower drain at 10 minutes is as follows:
K.A) fc Y ( KL/
=C,inexp ±— \+ —S- 1-exp 1-
'•m Q, J (,HC| ^ Q,
,-, nni /T ( 4.5 L/min^ f2.4x10 5mg/LV, ( 4.5 L/min^
C =0.01 mg/L» exp - . + s 1-exp =0.0090mg/L
i.out \ 9.1 L/miiJ I 0.0033 Jl I 9.1 L/minJJ
Step 8: Calculate mass emission rate as a function of time.
The mass emission rate for MEK at 10 minutes is calculated using Equation 2-32:
= Qi(Ci,in-CI>ouUomin) = 9.1 L/muiute • (0.01 mg/L - 0.0090 mg/L) = 0.0091 nig/min.
The resulting mass emission rate as a function of time is shown in Figure 8-1. The lower
value of KLA for MEK resulted in a significantly lower mass emission rate. The same eight-step
procedure was applied for toluene and DBCM, which resulted in a KLA value of 12 L/minute for
DBCM. The mass emission rate for DBCM is slightly lower than the rate for toluene in that
DBCM has a lower Henry's law constant than toluene.
8-6
-------
In previous modeling exercises, it has been assumed that the overall mass transfer coefficients
between two chemicals may be solely related by *Pj = KLi/KTj. This relationship requires only
knowledge of liquid molecular diffusion coefficients for each compound in accordance with
Equation 2.12, and is valid when gas-phase resistance to mass transfer is negligible for each
compound. As discussed previously, an assumption that gas-phase resistance is negligible is often
reasonable only when both compounds are highly volatile (e.g., cyclohexane and radon). Equation
2.15, used to predict *Pm, incorporates a chemical's liquid- and gas-phase resistance to mass
transfer and will converge to T, as kg/k, and/or Hc for both i and j become relatively large. Thus,
*Fm is a more appropriate value to predict values of KLA for chemicals of wide-ranging volatility.
However, in the case of showers, the value of kg/k, is sufficiently large that the value of KLA
for even chemicals as low in volatility as MEK may be estimated using T,. As a result, the more
important variable to predict is the chemical's Henry's law constant, which affects the
concentration driving force for mass transfer (Equation 2.28) and hence mass emission rates. For
this example, the emitted mass of DBCM was approximately 600 u.g, and the emitted mass of
MEK was approximately 120 ug.
8.2. DISHWASHER MODEL APPLICATION
Mass balance Equations 2-23 and 2-24 may be used to predict chemical emissions during a
dishwasher event of single or multiple cycles, that is, number of separate fills during operation.
For this example, the following dishwasher event was assumed: a prerinse cycle of 3.5 minutes, a
wash cycle of 1,0 minutes, and two rinse cycles of 6 and 14 minutes, respectively. Each cycle was
followed by a 2-minute drain period. The cycle order and times were based on those for the
experiment dishwasher. Other specific operating conditions included a water temperature of 55°C
and a liquid fill volume of 7.4 L resulting in a headspace volume of 181 L. Based on experiment
results, the headspace ventilation rate was assumed to be 5.7 L/minute. The value of KLA for
toluene (used in Equations 2-23 and 2-24 to predict liquid and gas-phase concentrations associated
with these operating conditions) was assumed to be 35 L/minute (average of KLA values
determined for Experiments 5 through 8 replicate in Table 5-5) for all cycles. Because of the
relatively small difference in values of KLA between experiments of different operating conditions
8-7
-------
(wash versus rinse), the value of KLA chosen for toluene represented the average of all dishwasher
results for heated water experiments (see Section 5.4.3).
The mass emission rate for toluene was predicted using the following steps:
Step 1: Predict liquid-and gas-phase concentrations as a function of time.
The liquid-phase concentration hi the dishwasher water after 3.5 minutes of operation for
the first cycle (prerinse cycle) is predicted using Equation 2-23:
c,=c,,
1.0
z z
C, = O.Olmg/L
exp -
5.0/min
mg/(L/min) 0.15/min • O.Olmg/L 5.0/min • O.Olmg/L
"" ' ' ' "f~ ..— . . .. —. —•' •"
4.7/min 4.7/min 2
1
^
(5.0 / min) :
--0.15/min
— 5.6
where
D = Z + Y = 4.7 /min + 0.34 /min = 5.0 /min
8-8
-------
z=KLA=35L/min
•V, 7.4L
v= QS KLA = 5.7L/min 35L/min = p 34/mm
VB VgHc 181L +181L»0.63
E = ZY-BX = 4.7/min • 0.34/min - 7.5/min • 0.19/min = 0.15/min
B = *iA= 35L/min = 7.5 /min
V,HC 7.4L«0.63
X =
_ K,A_ 35L/min _
Vg 181L
= zc60+xq0 = 4.
= 0.19/min
0 + 0.19/min • 0.01 mg/L = 0.0019 mg/(L • min)
The gas-phase concentration at the end of the first cycle is calculated using Equation 2-24:
Cs = Omgy^exp | - 5-I/mln 3 5min jcoshj j
.1/min)
--0.15 3.5min
+ 1.9xlO-3mg/(L»min)-
5.1/min»Omg/LN
- 0.
.
Step 2: Calculate the mass emission rate as a function of time.
Using Equation 2-31, the mass emission rate at the end of the first cycle is
8-9
-------
E = Qgcg,3.smin = 5.7 L/ininute • 3.5 x 10'4 mg/L = 0.0020 mg/min.
Step 3: Predict ventilation decay rate during drain period.
Between each cycle was a drain period, where water used during the cycle was pumped
from the machine. During the drain period, the gas phase was modeled using the
following equation with a ventilation rate of 5.7 L/minute:
(8-1)
where
Cg = headspace concentration (M/L3)
Cg0 = headspace concentration at end of cycle (M/L3)
Qg = machine ventilation rate (L3/T)
Vg = machine headspace volume (L3)
t = time (T).
The concentration of toluene in the dishwasher headspace at the end of the 2 minute drain
period is:
= 3.5xlO-4mg/L.
lol L
2minl = 3.3x 10-4mg/L.
J
Step 4: Repeat Steps 1 to 3 for number of dishwasher cycles.
Each cycle was modeled separately with an inlet liquid-phase concentration of 10 ug/L.
However, the gas-phase concentration of each cycle was dependent on that of the
previous cycle; that is, the initial gas-phase concentration for each cycle (Cg0) was equal
to the final gas-phase concentration of the previous drain cycle.
8-10
-------
The total mass of emitted toluene was calculated by integrating under the mass emission rate
curve shown in Figure 8-2. For this example, the total mass of toluene emitted over the entire
cycle was predicted to be 157 u.g. It should be noted that an additional 117 u.g of residual toluene
was retained in the dishwasher headspace at the end of the final rinse cycle. This residual would
be released as a "puff' if the dishwasher were opened soon after the final cycle. This more
concentrated release might contribute a greater exposure route than corresponding emissions
during the actual dishwasher operation. The stripping efficiency for toluene over all dishwasher
cycles was 93%.
By means of the first six steps outlined in Section 8.1, values of KLA for DBCM and MEK
were estimated. Although values of KLA were less important for this source because of
equilibrium limitations, a value of KLA for each chemical was needed to properly use the mass
balance model. As discussed in Section 5.4.4, a value of kg/k, for dishwashers was not determined.
Thus, to predict a value of KLA for a chemical of interest, a kg/k, ratio had to be assumed. Given
the hydrodynamic similarity between dishwashers and showers, the kg/k, value of 160 determined
for showers was used. The value of KLA estimated for DBCM was 37 L/minute, resulting in an
0
0 5
Key: D = Drain
10 15 20 25 30
Time (minutes)
35
40
45
0 Toluene + Dibromochloromethane n Methyl Ethyl Ketone
Figure 8-2. Mass emission rates for three chemicals for example dishwasher event.
8-11
-------
overall emitted mass of 143 fag, and stripping efficiency of 84%. As for toluene, a potential puff
release mass was calculated for DBCM to be 107 |j.g. The value of KLA for MEK for this example
was estimated to be 5.7 L/minute, resulting hi a total mass emitted of 2.5 jig, a puff release of 1.9
ug, and a stripping efficiency of 1.5%. Using the identical operating conditions listed for toluene
and an inlet concentration of 10 p,g/L yielded mass emission rates for each chemical as presented
in Figure 8-2. Again, the lower values of KLA and Henry's law constant for MEK resulted in a
significantly lower mass emission rate. To better see the shape of MEK emissions over time, the
ordinate of Figure 8-2 was magnified as shown in Figure 8-3.
The general shape of the mass emission rate curve reflected the approach to a dynamic
equilibrium condition for each chemical. Although DBCM had a slightly greater value of KLA
than toluene, the mass emission rate for DBCM was lower because of equilibrium limitations in
the headspace. Thus, for equilibrium-limited cases, the value of KLA for a given chemical merely
indicates how rapidly equilibrium will be achieved within the headspace. As a result of the
insignificance of KLA, more emphasis is placed on the accuracy of a chemical's Henry's law
8 E-05-
.9 6.E-05
•5*
M
•I 4.E-05-
.2
i
CO
tfl
S 2.E-05-
o F+not
(
U-i^U^g ^y ^ <^
D
> 5 10 15 20 25 30 35 40 4
Time (minutes)
5
Figure 8-3. Amplification of Figure 8-2 to show MEK mass emission rate.
8-12
-------
constant. Currently, there is a lack of information regarding Henry's law constants for potential
drinking water contaminants, especially at higher temperatures.
8.3. WASHING MACHINE MODEL APPLICATION
Different mass balance equations were used to predict emissions from each washing machine
cycle. Mass balance Equations 3-8 and 3-9 were used to predict chemical concentrations .during
the fill cycle of a washing machine event. Mass balance Equations 2-23 and 2-24 were used to
predict chemical concentrations during the wash and rinse cycles of a washing machine event.
Similar to dishwashers, each cycle was modeled separately, with the initial conditions reflecting
previous cycles. For example, the initial liquid- and gas-phase concentrations for the wash cycle
were equal to the final liquid- and gas-phase concentrations, respectively, for the first fill cycle.
Both fill cycles had an identical inlet chemical liquid-phase concentration of 10 ug/L. As for
dishwashers, the headspace concentration during the drain/spin period for a washing machine was
modeled using Equation 8-1 and the emission rate was calculated using Equation 2-31.
For this example, a washing machine event was assumed to consist ofa 3.3-minute fill cycle
at a flowrate of 13.8 L/minute (« 46 L total liquid volume), a 10-minute wash cycle, a 4-minute
drain and spin cycle, another 3.3-minute fill cycle also at a flowrate of 13.8 L/minute, a 4-minute
rinse cycle, and finally a 6-minute drain and spin cycle. Other specific operating conditions for
this example were a water temperature of 21°C, and ventilation rates of 55 L/minute for the fill
cycle, and 53 L/minute for the remaining cycles. With a fill volume of 46 L and an approximate
equivalent clothing volume of 11 L, the headspace volume was 92 L.
To predict mass emissions associated with the example operating conditions, a value of KLA
for toluene was chosen for each cycle. An "average" value of KLA was not used for all cycles
because of the significant effects of operating conditions on KLA observed for washing machine
experiments. Values of KLA for toluene were as follows: 2.9 L/minute for both fill cycles, 0.58
L/minute for the wash cycle, and 0.84 L/minute for the rinse cycle. On the basis of these values of
KLA and an inlet concentration of 10 ug/L, the mass emission rate was calculated using the
following steps:
8-13
-------
Step 1: Predict liquid- and gas-phase concentrations as a function of time for fill cycle.
A second-order Runge-Kutta solution technique was used to determine the liquid- and
gas-phase concentrations during filling. The applicable general second-order solution
technique is:
xn+1 =xn +—{f(tn,xn)+f[tn +At,xn +f(tn,xn)]
(8-2)
Applying this method to Equations 2-25 and 2-26 and using 1-second time steps enabled
prediction of the liquid- and gas-phase concentrations at each time step, respectively.
The liquid-phase concentration in the washing machine water after filling for 3.3 minutes
follows:
Find first-order solution:
vn+1 —
A. ~~
3,Cl|h Q.C," KLAC,n | KLACe"
V,n V," V,"
V,"HC
n+l _ri3.8L/min»0.01mg/L 13.8L/min*0.0084mg/L 2.9L/min»0.0084mg/L 2.9L/min»3.5xIQ^mg/Ll
~[ 45.8L 45.8L ~ 45.8L + 45.8 L« 0.24 J
• (1 / 60 min+0.0084 mg / L
Findf(tn,xn):
f(f .cn)=
V,"
Q.C," KLAC," | KLACg"
V," V,"HC
8-14
-------
fCt" x"\= 13 -8L /min • 0-01 mg /L 13 .8 L/ min •0.0084 mg / L
^ 'X '~ 45 .8L 45 .8L
2.9 L /min • 0.0084 mg / L 2.9 L / min • 3.4 x 10 ~4 mg / L
-- -ITil - + - 45.8L.0.24
„ „ , ,
= 3.0xlO-*mg/(L.nun)
Find f[tn+At,xn+Atf(tn,xn)]:
f(t" + At,x«
vn+l W n+'
Vl Vi
f(tn +At x" +Atfft" x"))^ 13.8L/min«0.01mg/L 13.8L/miti«0.0084mg/L
^ 'X . ^ 'X "~ 46L 46L
2.9L/min»0.0084mg/L 2.9L/min»3.4xl(r4mg/L
•-.--. 46L + 46L.0.24
Insert appropriate values into Equation 8-2:
II 6.0 min
C,n+I =0.0084 mg/L + :-^^--(3.1xlO-5mg/L«min+3.0xlO-5mg/L«min)=0.0084mg/L
Note: The second-order solution is virtually equivalent to the first-order solution; thus,
a fourth-order solution technique was not deemed necessary. Also, values used in this
example were rounded. More exact values were used in spreadsheet calculations.
Similarly, the gas-phase concentration in the washing machine headspace at the end of
filling is calculated as follows:
8-15
-------
Find first-order solution:
Cg-':
.-v,') (vt-v,n) (vt-v,"j
C n*i_['-5SL/min»3.4xlO"4mg/L t 13.8L/min«3.4xlQ^mg/L 2.9L/min« 0.0084mg/L 2.9L/min«3.4xlO"4mg/Ll
1 L (150L-45.8L) (150L-45.8L) + (150L-45.8L) (l SOL-45.8L)» 0.24 J
• (1/60 min)+3.4X10"4 mg/L = 3.5xlO-5mg/L
Findf(tn,xn):
f(t",xn)=
Q.C8" . KLAC,n KLACgn
-v,n) (vt-v,") (vt-v,n) (v.-v.-
ftn xn
^ >X '
(150L-45.8L)
13..8L./ min* 3.4x10"* mg/L
(150L-45.8L)
+ 2-9L/min • 0.0084 mg / L 2.9 L /min • 3 .4x10 ~4 mg /L
(150L-45.8L) (150 L - 45 .8 L)* 0.24
Find
f(t" +At,x" +Atf(tn,xn)) =
_(vt -v>+1J+ (v,-v,n+1 J+(vt-v;
KLAC," KLACgn
n+1
fft" +At x" +AtfCtn :cnV\_-55L/min»3.5xlQ-4mg/L O.SL/min^S.SxlO^mg/L
V ' V ' " (150L-46L) (150L-46L)
+ 2-9L/min "0.0084 mg / L 2.9 L/min «3.5xlQ-4mg /L
(150 L - 46 L) (150 L - 46 L)« 0.24
8-16
-------
Insert appropriate values into Equation 8-2:
C gn + 1 = 3 .5x10 -" mg/L +
"60
(5.6x10 5 mg/L «> min + 5 .6 x 10 ~5 mg/L • min )
Step 2: Calculate mass emission rate for each time step during fill cycle.
Through use of Equation 2-31, the mass emission rate at the end of the fill cycle is:
E = QgCgj3.4min - 55 L/minute • 3.5 x 1Q-4 mg/L = 0.019 mg/min.
Step 3: Predict liquid- and gas-phase concentrations as a function of time during wash
cycle.
Equations 2-23 and 2-24 were used to predict liquid- and gas-phase concentrations as a
function of time. Refer to dishwasher steps for use of equations. The initial liquid-
phase concentration is equal to the final fill liquid-phase concentration, which in this
example is 0.0084 mg/L. Likewise, the initial gas-phase concentration is equal to the
final fill gas-phase concentration, which in this example is 3.5 x 10'4 mg/L. At the end
of the 10-minute wash cycle, the estimated liquid-phase concentration is 0.0075 mg/L
and the estimated gas-phase concentration is 8.1 x 10'5 mg/L.
Step 4: Calculate mass emission rate as a function of time for wash cycle.
Again, with the use of Equation 2-31, the mass emission rate may be calculated. For
this example, the rate is:
E = 53 L/minute • 8.1 x 1Q-5 mg/L = 0.0043 mg/min.
Step 5: Predict ventilation decay rate during drain period.
Between the end of the wash cycle and the next fill is a drain period, where water used
during the wash cycle is pumped from the washing machine. This drain/spin cycle was
8-17
-------
modeled using Equation 8.1 with a ventilation rate of 53 L/minute. The concentration
of toluene in the washing machine headspace at the end of the 4-minute drain period is
nearly zero.
Step 6: Repeat Steps 1 through 5 for rinse fill, rinse, and final drain.
The wash and rinse cycles were modeled separately, both with an inlet liquid-phase
concentration of 10 u.g/L.
The total mass emitted for the entire washing machine event was 210 jj,g. The mass
emission rate is shown in Figure 8-4. The mass of toluene remaining in the headspace after the
final spin cycle was 0.41 u,g, significantly lower than the residual mass observed in the
dishwasher headspace. The low residual washing machine headspace mass may be attributed to
its relatively high ventilation rate, which effectively flushes the headspace of the machine. The
stripping efficiency integrated over all cycles for toluene was 22%.
10 15 20 25
Time (minutes)
30
35
o Toluene + Dibromochloromethane n Methyl Ethyl Ketone
Figure 8-4. Mass emission rates for three chemicals for example washing machine event.
8-18
-------
In addition to having different values of KLA, each type of cycle was characterized by a
different kg/k, ratio. The kg/k, values chosen for this example were 9.5 for the fill cycles and 2.2
for the wash and rinse cycles. The resulting values of KLA for DBCM using the six-step
procedure described in Section 8.1 were 1.1 L/minute for the fill cycles, 0.12 L/minute for the
wash cycle, and 0.18 L/minute for the rinse cycle. The resulting values of KLA for MEK
following the same procedure were 0.31 L/minute for the fill cycle, 0.030 L/minute for the wash
cycle, and 0.044 L/minute for the rinse cycle.
To illustrate the importance of gas-phase resistance to mass transfer, the total mass
emissions for DBCM and MEK were calculated using values of KLA based on Tm and values of
KLA based only on XF1. The total mass emitted for DBCM using Ym to predict KLA was 67 jag
(stripping efficiency of 7.1%) compared withl50 |j,g emitted when ¥, was used to predict KLA.
The total mass emitted for MEK using »Fm to predict KLA was 18 jag (stripping efficiency of
1.9%) compared with 65 u.g emitted when Y, was used to predict KLA.
8.4. BATHTUB MODEL APPLICATION
The same mass balance equations used for modeling emissions from washing machines
were used for bathtubs (see washing machine steps). Equations 3.8 and 3.9 were used to predict
chemical concentrations in the liquid and gas phases, respectively, during the fill portion of
bathtub use. Equations 2.23 and 2.24 were used to predict liquid- and gas-phase concentrations
during the bathing portion of bathtub use. Equation 2.33 was used to predict resulting mass
emissions. The inlet chemical concentration was 10 ug/L. The initial concentrations for the
bathing portion were equal to the final liquid- and gas-phase concentrations for the fill portion.
For this example, a bathtub was assumed to be filled for 8 minutes using a water flowrate of 9.1
L/minute, resulting in a total liquid volume of approximately 73 L. There was a 20-minute
bathing period after the filling experiment. It was also assumed that the bathing event occurred
in a 13 m3 bathroom with an air exchange rate of 1.0/hour (Qg = 217 L/minute). The temperature
of the water was equivalent to the warmest experimental temperature of 36°C and remained
constant for the entire bathing event.
8-19
-------
The values of KLA chosen for toluene were based on bathtub fill and surface volatilization
experiments, respectively. Given the narrow range of bathtub fill results, an average value of 4.4
L/minute was chosen based on high flowrate average. Similarly, an average value of 1.2
L/minute was chosen to represent surface volatilization with a person present. The resulting
mass emissions for the entire bathtub event are plotted hi Figure 8-5. Integrating under the mass
emission rate curve, the total mass of toluene emitted was 375 jag, with a corresponding
integrated stripping efficiency of 51% .
As with the previous sources, values of KLA were predicted for DBCM and MEK. For a
bathtub event, the kg/k, ratio associated with filling was chosen to be 51 and the kg/k] ratio
associated with bathing was chosen to be 70. The values of KLA estimated for DBCM using ^¥m
were 4.0 L/minute for filling the tub and 1.1 L/minute for surface volatilization. The values of
K^A estimated for MEK using ¥„, were 0.72 L/minute for filling the tub and 0.25 L/minute for
surface volatilization. The mass emission rates for these two chemicals are presented in Figure
8-5.
The total mass emitted of DBCM using ^¥m to predict KLA was 350 u.g (stripping efficiency
of 48%) compared with 380 u.g emitted using T, to predict KLA. The total mass emitted of MEK
using ¥m to predict KLA was 89 jag (stripping efficiency of 12%) compared with 250 |j,g emitted
using ^ to predict KLA.
fl
1 0.030 .
«§• 0.025 -
o
| 0.020 -
1 0.015 -
CO
| 0.010 -
J 0.005 -
1=5 0.000 -
(
Fill
--
Bathing
nTTTWWWWWi^
)
5 10 15 20 25
Time (minutes)
o Toluene + Dibromochloromethane Q Methyl Ethyl Ketone
3
0
Figure 8-5. Mass emission rates for three chemicals for example bathtub event.
8-20
-------
9. SUMMARY AND CONCLUSIONS
9.1. SUMMARY
The research described in this report consisted of 113 experiments involving 5 tracer
chemicals (acetone, ethyl acetate, toluene, ethylbenzene, and cyclohexane) and 4 sources
(showers, dishwashers, washing machines, and bathtubs). The source experiments completed for
this study hav$ significantly increased the existing knowledge base of published experiments
involving chemical volatilization from drinking water to indoor air.
A determined attempt was made to follow a rigorous quality assurance project plan and to
perform mass closure assessments on all sources. The latter are not often reported in the
published literature and necessitate simultaneous collection of liquid and gas samples. The
resulting database allowed determination of chemical stripping efficiencies (r|), overall mass
transfer coefficients (KLA), and liquid- and gas-phase mass transfer coefficients (k]A and kgA)
for each chemical and source. In addition, headspace ventilation rates were measured for
washing machines and dishwashers.
An important contribution from these experiments was the calculation of mass transfer
parameters for acetone and ethyl acetate, two chemicals with Henry's law constants considerably
lower than any used in previous experiments. Use of these lower volatility chemicals greatly
improves the ability to estimate mass transfer coefficients for many disinfection by-products,
pesticides, and other low-volatility compounds.
The ratio of gas- and liquid-phase mass transfer coefficients-(kg/k,) was calculated for
varying operating conditions for each source. As described in Chapter 2 and illustrated in
Chapter 8 of this report, knowledge of the magnitude of kg/k, is critical for improved estimates of
chemical volatilization from drinking water to indoor air. This is particularly true when an
overall mass transfer coefficient for one chemical is to be extrapolated to a second chemical of
interest. Experimentally predicted values of kg/k[ should serve as a major contribution of this
study.
9-1
-------
A series of dynamic emission models were developed for each source and were based on
fundamental reactor analyses, mass balances, and mass transfer kinetics. The experimental mass
transfer coefficients, air exchange rates, and protocols described in this report can be used as
direct input values or to estimate reasonable input values for the reported emission models. In
addition to input parameters based on this research, values based on previous research are
provided. All experimental results related to chemical volatilization from tap water are now
available in a single database provided in the Appendix. The database contains 454
experimental results (including 293 results from this study) and is a valuable source of
information for those interested in modeling human inhalation exposures related to specific
contamination scenarios.
Values of KLA, kgA, kA, kg/k,, and headspace ventilation rates are presented in Chapters 4
through 7. A summary of chemical stripping efficiencies is provided in Table 9-1 for each
source. The ranges of stripping efficiencies for each chemical and source reflect the fact that, for
some sources, the rate of volatilization from drinking water to indoor air, and subsequent human
exposure, is highly dependent on source operating conditions; In many cases, assuming 100%
volatilization will significantly overestimate emissions and human inhalation exposure to
chemicals originating in drinking water.
A set of general conclusions is provided in Section 9.2. Source-specific conclusions are
provided in Sections 9.3 through 9.6. Finally, recommendations for future research are provided
in Section 9.7.
Table 9-1. Summary of experimental stripping efficiencies and kg/k,
Chemical
Acetone
Ethyl Acetate
Toluene
Ethylbenzene
Cyclohexane
kA
Showers
6.3-16
15-36
6.1 - 77
62-75
65-80
110-223
Bathtubs3
2.6 - 14
4.6-16
35-53
33-54
64-69
Flow-through: 37-96
Fill: 27 - 77
Bathing: 54 - 78
Dishwashers
18 - 55
*
96-98
97-98
100
*
Washing machines'*
3.8-38
*
30-99
31-99
40 - 100
Fill: 4.5-20
Wash/Rinse: 0.13-8.6
'Stripping efficiency based on combined effects of filling and bathing (20 minutes) in series.
Stripping efficiency based on combined effects of fill and wash (or rinse) in series.
*Unable to determine
9-2
-------
9.2. CONCLUSIONS: GENERAL
The findings of this study lead to several genera! conclusions about all, or most, sources and
chemicals:
1. System operating conditions can have a significant effect on chemical emissions. This is
particularly true for lower volatility chemicals (e.g., ethyl acetate and acetone) and for all
chemicals emitted from washing machines.
2. For higher volatility chemicals (e.g., with Henry's law constants greater than toluene),
chemical stripping efficiencies should vary by approximately 30% or less (relative) over a
wide range of system operating conditions for showers, bathtubs, and dishwashers.
3. Although many operating conditions can affect chemical volatilization rates from water to
indoor air, water temperature appears to have the greatest effect over all sources and
chemicals. This is generally caused by a combination of increases in Henry's law constants
(Hc) and liquid-phase mass transfer coefficients, as well as by increases in headspace
ventilation rates for washing machines and dishwashers.
4. Chemical stripping efficiencies increase as Henry's law constant increases from lower
volatility chemicals, (e.g., acetone and ethyl acetate) to higher volatility chemicals (e.g.,
toluene, ethylbenzene, and cyclohexane). However, stripping efficiencies are relatively
insensitive to Henry's law constant for Hc greater than that of toluene. The one exception is
for the fill cycle of bathtubs.
5. Failure to properly account for gas-phase resistance to mass transfer can lead to significant
overestimates of chemical volatilization to indoor air. This is particularly true for lower
volatility chemicals or those sources with low values of gas- to liquid-phase mass transfer
coefficients (kg/k,)(e.g., washing machines).
9-3
-------
6. Significant improvements in emissions estimates are possible by incorporating appropriate
values of kg/k, into the estimation process. However, values of kg/k, are specific to
individual systems and operating conditions and should therefore be chosen carefully.
7. The use of dual tracers with similar physicochemical properties (e.g., toluene and
ethylbenzene) can be effective as internal checks of the quality of experimental results.
9.3. CONCLUSIONS: SHOWERS
Showers are the most extensively studied source of chemical volatilization from drinking
water to indoor air. However, previous studies are limited in terms of the separation of gas- and
liquid-phase mass transfer coefficients, and analyses of compounds with significant gas-phase
resistance to mass transfer. Furthermore, previous researchers have speculated on the validity of
the assumptions that shower stall atmospheres are well mixed. The following conclusions stem
from this study and directly address the issues listed above:
1. The shower stall atmosphere is relatively well mixed; that is, gas-phase concentrations
should be relatively homogenous throughout a shower stall.
2, Gas-phase resistance to mass transfer dominates the overall resistance to mass transfer for
acetone and chemicals with Henry's law constants lower than Hc for acetone. Gas- and
liquid-phase resistances are both important for compounds with Hc similar to ethyl acetate.
3. Liquid-phase resistance to mass transfer is dominant for chemicals with Hc greater than
toluene. Thus, overall mass transfer coefficients for two or more of these compounds can
be related using only liquid-phase molecular diffusion coefficients.
4. On the basis of the results of this study, stripping efficiencies for chemicals with Hc greater
than or equal to toluene should range from 60% to 80%. Slight deviations between this
9-4
-------
range and those previously reported by other researchers are likely caused by differences in
showerheads.
5. Mass emission rates from water to shower stall air decrease with time depending on the
volatility of a chemical and its approach to equilibrium. Thus, knowing a chemical's
Henry's law constant at typical shower temperatures is important.
9.4. CONCLUSIONS: DISHWASHERS
Previous to this study, dishwashers had received little attention as emission sources. The
following conclusions are derived from 29 experiments involving a commercial dishwasher:
1. Dishwashers are characterized by a low, but continuous and constant, headspace ventilation
rate during operation.
2. Dishwashers are extremely effective at stripping a wide range of chemicals from water.
3. High stripping efficiencies (approaching 100% for chemicals with Hc > toluene) can be
attributed to the high water temperatures used in dishwashers, relatively high liquid-phase
mass transfer coefficients associated with airborne droplets, and large headspace volume
relative to water volume within the dishwasher.
4. Chemicals are rapidly stripped from water to the interior headspace of dishwashers during
the first minute of water spraying, that is, during rinse and wash cycles. Thereafter,
chemicals with Henry's law constants less than or equal to those associated with
ethylbenzene approach a state of chemical equilibrium.
5. The most significant release of chemicals to indoor air would occur if the dishwasher door
is opened immediately after use.
9-5
-------
9.5. CONCLUSIONS: WASHING MACHINES
Chemical emissions from washing machines have received some attention in the published
literature. However, variations in machine operating conditions and chemical properties have
been limited. Furthermore, previous researchers have not addressed the potential for emissions
during fill cycles. The following conclusions come from 57 experiments using a commercial
washing machine:
1. Washing machines are well ventilated, with air exchange rates exceeding those for
dishwashers by one to two orders of magnitude.
2. The use of hot water leads to significant increases in washing machine ventilation rates
caused by buoyancy-induced air flows.
3. Stripping efficiencies from washing machines, more than any other source, are extremely
sensitive to system operating conditions.
4. Under appropriate conditions (hot water, rinse cycle with no detergent present, low clothes
loading), stripping efficiencies for chemicals with Henry's law constants greater than
toluene can approach 100%.
5. Chemical emissions during machine filling are generally lower than emissions during wash
and rinse cycles.
6. During wash and rinse cycles, gas-phase resistance to mass transfer is important for all
chemicals with Henry's law constants less than or equal to Hc for ethylbenzene.
9.6. CONCLUSIONS: BATHTUBS
Previous research has focused on volatilization of chemicals during shower, events. The
argument for such studies is that, unlike other sources, showers lead to potentially significant
9-6
-------
human exposure. However, this is also true for bathing, that is, volatilization from bathtubs to
indoor air. For this study, 17 experiments were completed to assess chemical volatilization
during the filling of bathtubs, during bathing events, and for applications in which water enters a
tub and is allowed to flow directly into an open drain. Significant conclusions are listed below:
1 . Chemical stripping efficiencies are similar during flow-through, filling, and bathing events.
Here, bathing refers to volatilization from the water surface with mild agitation of the water
over a 20-minute event.
2. Stripping efficiencies are more sensitive to chemical properties, particularly Henry's law ,
constant (Hc), during tub filling than during flow-through or surface volatilization events.
This is because of the significant degree of air entrainment that occurs as the water jet
impacts the underlying bath pool. Diffused air bubbles promote gas-liquid mass transfer,
that is, chemicals transferred from liquid to bubbles. This transfer mechanism is sensitive to
Hc, as accumulation of chemical mass in bubbles can lead to an approach to chemical
equilibrium for compounds with low Hc. Furthermore, gas-phase resistance to mass transfer
tends to be more significant when air bubbles are present. This fact is consistent with the
relatively low value of kg/k] for fill events.
3. Combined stripping efficiencies for fill and bathing events are slightly lower, but
comparable, to those associated with shower events. Not included in the bathtub estimate
are emissions from water flowing through the tub when attempting to attain a water
temperature appropriate for bathing.
4. Because of the longer exposure tunes, chemical emissions during the use of bathtubs may
be as, or more, significant than during showers, in terms of human inhalation. This is
particularly important given that small children are typically washed in bathtubs rather than
showers and are generally more sensitive to chemical exposure than are healthy adults.
9-7
-------
9.7. RECOMMENDATIONS FOR FUTURE RESEARCH
The research described herein has increased the existing knowledge base associated with
chemical volatilization from water to indoor air. A natural progression would be to use the
results of this study to complete an integrated exposure assessment for various water
contaminants and water usage scenarios. The assessment should include bounds that account for
uncertainties in mass transfer parameters and other relevant exposure factors. However, based
solely on tibe findings of the study described in this report, recommendations for future research
are as follows:
1. The database that now exists for volatilization during showers is large. Additional research
related to chemical volatilization during showers is not recommended. However,
information regarding aerosol formation and the associated impact on chemical
volatilization for showers and other relevant sources is warranted.
2. Currently, there is a lack of information on Henry's law constants at elevated temperatures
for most potential drinking water contaminants. As shown by this work, the total mass
emissions from several sources are highly dependent on this parameter and could be
significantly over- or underestimated for these chemicals. Additional research is needed to
determine Henry's law constants at water temperatures ranging from 30°C to 60°C.
3. The results of this study suggest that volatilization during the use of bathtubs may be as, or
more, significant than during showers in terms of human exposure, particularly for small
children. Additional research may be warranted to improve estimates of chemical
volatilization from bathtubs, including an assessment of the effects of soap films on
retardation of mass transfer. Furthermore, additional studies should be completed to better
simulate me kinetic energy imparted on bath water by human bathing activities.
4. The database that has been generated for, dishwashers is novel and suggests that chemical
equilibrium is achieved rapidly between the water and the dishwasher headspace. As such,
9-8
-------
the rate of air exchange between dishwasher and room atmospheres is a critical parameter.
In this study, that rate was quantified for a commonly purchased commercial dishwasher. It
would be beneficial to repeat a series of air exchange experiments on several commercially
available dishwashers. These experiments could be used to quantify a range or distribution
of air exchange rates. ,
5. Washing machines may need the greatest amount of additional research based on the range
of experimental results achieved for a single washing machine, and the potential for nearly
complete volatilization of many chemicals. Specific research should focus on improved
estimates of headspace air exchange rates, particularly during the use of hot water, and the
formation of chlorination by-products during the use of sodium-hypochlorite containing
bleaches. The latter was not studied as part of this research effort, but was previously
documented by the principal investigator of this project. Additional washing machine
designs should be considered, including upright machines (vertical door) such as are often
used in laundromats.
6. Several additional sources of chemical volatilization from water to indoor air were not
studied during this research effort and are deserving of future experimental work. These
include toilets, indoor saunas and tubs, aquariums, and humidifiers. The latter should
include centralized humidification systems, which are now common in many parts of the
United States.
7. It would be beneficial to incorporate the source models developed herein into an indoor air
quality model.
8. The model described in Recommendation 7 should be used in conjunction with information
related to water usage and human activity patterns to assess the significance of human
inhalation exposure to chemicals that originate in drinking water. An important example
9-9
-------
involves lower volatility disinfection by-products, which should be characterized by
significant gas-phase resistance to mass transfer.
9. The database and models developed during this research effort should be supported with
field data. In addition to water and air sampling in a home with a contaminated water
supply, other potential sources include laundromats, gymnasium shower facilities, and
restaurant kitchens.
9-10
-------
10. REFERENCES
Ashworth, RA; Howe, GB; Mullins, ME; Rogers, TN. (1988) Air-water partitioning coefficients of organics in
dilute aqueous solutions. J Haz Mat 18:25-36.
Bernhardt, G; Hess, CT. (1995) Acute exposure from radon-222 and aerosols in drinking water. Written
communication. Unpublished master's thesis. University of Maine.
Box, GEP; Bisgaard, S. (1988) Statistical tools for improving designs. MechEng 110:32-40.
Corsi, RL; Howard, C; Strader, R. (1996) Volatilization rates from water to indoor air: phase I report. Project
report submitted to U.S. Environmental Protection Agency.
Danckwerts, PV. (.1951) Significance of liquid film coefficients in gas absorption. IndEngChem 43(6): 1460-1467.
Dobbins, WE. (1956) Mechanism of gas absorption by turbulent liquids. In: Advances in water pollution research
2. Eckenfelder, WW, ed. New York: Pergamon Press, Ltd., pp. 61-96.
Enviromega. (1993) Collection system organic release algorithm: user's manual, p. 13.
Gesell, TF; Prichard, HM. (1980) The contribution of radon in tap water to indoor radon concentrations. In:
Natural radiation environment III, Vol. 2. Houston: Technical Information Center, U.S. Department of
Energy, pp. 1347-1363.
Giardino, NJ; Andelman, JB. (1991) Poster paper presented at Annual Conference of the American Water Works
Association, Philadelphia, PA.
Giardino, NJ; Andelman, JB. (1996) Characterization of the emissions of trichloroethylene, chloroform, and 1,2-
dibromo-3-chloropropane in a full-size, experimental shower. J Exposure Anal Environ Epidemiol 6(4):413-
423. ......
Giardino, NJ; Hageman, JP. (1996) Pilot study of radon volatilization from showers with implications for dose.
Environ SciTechnol 30(4): 1242-1244.
Giardino, NJ; Andelman, JB; Borrazzo, JE; Davidson, CI. (1988) Sulfur hexafluoride as a surrogate for
volatilization of organics from indoor water uses. J Air Pollut Control Assoc 38(3):278-280.
10-1
-------
Giardino, NJ; Esmen, NA; Andelman, JB. (1992) Modeling volatilization of trichloroethylene from a domestic
shower spray: the role of drop-size distribution. Environ Sci Technol 26:1602-1606.
Hess, CT, Weiffenbach, CV; Norton, SA. (1982) Variations of airborne and waterborne Rn-222 in houses in
Maine. Environ Int 8:59-66.
Higbie, R. (1935) The rate of exposure of a pure gas into a still liquid during short periods of exposure. Trans Am
Inst Chem Eng 31:365-388.
Hodgson, AT; Garbesi, K; Sextro, RG; Daisey, JM. (1988) Evaluation of soil-gas transport of organic chemicals
into residential buildings: final report. Lawrence Berkeley Laboratory Report, contract no. DE-AC03-
76SF00098.
Hopke, PK; Raunemaa, T; Datye, V; Kuuspalo, K; Jensen, B. (1995) Assessment of exposure to radon and its
decay products from showering in radon-laden water. In: Indoor air: an integrated approach. Morawska, L;
Bofinger, ND; Maroni, N, eds. Oxford: Elsevier Science Ltd., pp. 107-110.
Howard, CL. (1998) Volatilization rates of chemicals from drinking water to indoor air. Ph.D. dissertation, The
University of Texas at Austin.
Howard, PH, ed. (1989) Handbook of environmental fate and exposure data for organic chemicals (vol. I). Boca
Raton, FL: CRC Press, Inc.
Howard, PH, ed. (1990) Handbook of environmental fate and exposure data for organic chemicals (vol. II).
Chelsea, MI: Lewis Publishers.
Howard-Reed, C; Corsi, RL; Moya, J. (1999) Mass transfer of volatile organic compounds from drinking water to
indoor air: the role of residential dishwashers. Environ Sci Technol 33:2266-2272.
Howard, C; Corsi, RL. (1996) Volatilization of chemicals from drinking water to indoor air: role of the kitchen sink.
J Air Waste Manage Assoc 46:830-837.
Howard, C; Corsi, RL. (1998) Volatilization of chemicals from drinking water to indoor air: the role of residential
washing machines. J Air Waste Manage Assoc 48:907.
Hsieh, CC; Ro, KS; Stenstrom, MK. (1991) Estimating stripping rates and gas/liquid mass transfer coefficients of
semi-volatile organic compounds in surface aeration. Proceedings of the 64th annual WPCF conference;
Toronto, Ontario.
10-2
-------
Hsieh, CC; Babcock, RW, Jr.; Stenstrom, MK. (1994) Estimating semi-volatile organic compound emission rates
and oxygen transfer coefficients in diffused aeration. Water Environ Res 66(3):206-210.
Jo, WK; Weisel, CP; Lioy, PJ. (1990) Routes of chloroform exposure and body burden from showering with
chlorinated tap water. Risk Anal 10(4):575-580. '
Keating, GA; McKone, TE. (1993) Measurements and evaluation of the water-to-air transfer and air concentration
for trichloroethylene in a shower chamber. In: Modeling of indoor air quality and exposure, ASTM STP 1205.
Nagda, NL, ed. Philadelphia: American Society of Testing and Materials, pp. 14-24.
Keating, GA; McKone, TE; Gillet, JW. (1997) Measured and estimated air concentrations of chloroform in
showers: effects of water temperature and aerosols. Atmos Environ 31(2): 123-130.
Lewis, WK; Whitman, WG. (1924) Ind Eng Chem 16:1215-1220.
Lide, DR, ed. (1995) CRC handbook of chemistry and physics (76th ed.) Boca Raton, FL: CRC Press.
Little, JC. (1992) Applying the two-resistance theory to contaminant volatilization in showers. Environ Sci
Technol 26(7): 1341-1349.
Mackay, D; Shiu, WY; Sutherland, RP. (1979). Determination of air-water Henry's law constants for hydrophobic
pollutants. Environ Sci Technol 13(9):333-337.
Matter-Muller, C; Gujer, W; Giger, W. (1981) Transfer of volatile substances from water to the atmosphere. Water
Res 15:1271-1279.
McKone, TE. (1987) Human exposure to volatile organic compounds in household tap water: the inhalation
pathway. Environ Sci Technol 21:1194-1201.
McKone, TE; Knezovich, JP. (1991) The transfer of trichloroethylene (TCE) from a shower to indoor air:
experimental measurements and their implications. J Air Waste Manage Assoc 41:832-837.
Moya, J; Howard-Reed, C; Corsi, RL. (1999) Volatilization of chemicals from tap water to indoor air from
contaminated water used for showering. Environ Sci Technol 33:2321-2327.
10-3
-------
Munz, C; Roberts, PV. (1989) Gas- and liquid-phase mass transfer resistances of organic compounds during
mechanical surface aeration. Water Res 23(5):589-601.
Partridge, JE; Horton, TR; Sensintaffer, EL. (1979) A study of 222Rn released from water during typical household
activities. Office of Radiation Programs, Eastern Environmental Radiation Facility, Technical Note
ORP/EERF-79-1. Montgomery, AL: U.S. EPA.
Roberts, PV; Dandliker, P; Matter-Muller, C. (1984) Volatilization of organic pollutants in wastewater treatment
model studies (EPA/600/2-84-047). Cincinnati: U.S. Environmental Protection Agency.
Schoene, K; Steinhanses, J. (1985) Determination of Henry's law constant by automated head space-gas
chromatography. Fresenius Z Anal Chem 321:53 8-543.
Shepherd, J; Kemp, J; Corsi, RL. (1996) Chloroform in indoor air and wastewater: the role of residential washing
machines. J Air Waste Manage Assoc 46(7):631-642.
Smith, JH; Bomberger, DC, Jr; Haynes, DL. (1980) Prediction of the volatilization rates of high-volatility
chemicals from natural water bodies. Environ Sci Technol 14(11): 1332-1337.
Tancrede, MV; Yanagisawa, Y; Wilson, R. (1992) Volatilization of volatile organic compounds from showers. I:
analytical method and quantitative assessment. Atmos Environ 26A(6): 1103-1 111.
Tucker, WA; Nelken, LH. (1990) Diffusion coefficients in air and water. In: Handbook of chemical property
estimation methods. Lyman, WH; Reehl, WF; Rosenblatt, DH, eds. Washington, DC: American Chemical
Society, pp. 17-1-17-25. ,
Wooley, J; Nazaroff, WW; Hodgson, AT. (1990) Release of ethanol to the atmosphere during use of consumer
cleaning products. J Air Waste Manage Assoc 40:1114-1120.
10-4
-------
APPENDIX:
CHEMICAL VOLATILIZATION DATABASE
A-l
-------
Equations to Solve Mass Transfer Parameters In Database
GonoralgquaUons Applied to All Appllcablo Sources:
Chemical Stripping Efficiency ~
Variables Units
C| = chemical concentration in liquid phase out of system mg/L
C^in = chemical concentration in liquid phase entering system mg/L
VgC^+Qjc^dt
Mass Closure Estimate -
Variables Units
V| = liquid volume L
CLI » chemical concentration In liquid phase at time 1 mg/L
Ci^ = chemical concentration in liquid phase at time 2 mg/L
Vg = headspace volume L
C0>1 = chemical concentration in gas phase at time 1 mg/L
Cg,2 = chemical concentration in gas phase at time 2 mg/L
Qg = ventilation rate of system L/mln
ti = time 1 min
12 = time 2 min
k^fk, Matrix Method
KiA Chemical 1 Chemical2 Chemicals Chemical4 Chemical n
Chemical 1
Chemical 2
Cfuntkaia
1
Variables
KLAi = overall mass transfer value for chemical "i"
Kj_Aj = overall mass transfer value for chemical "j"
D,=
«/ =
TS
D01-
Thrco n x n matrices were filled with the following values:
Matrix 1: Ratio of measured KiA/KLAt values for all chemicals and single experimental condition
Matrix 2: Ratio of predicted KiAj/KiA values for all chemicals using an assumed kg/k| value
In the following equation:
Hds
H^=
liquid-phase diffusion coefficient for chemical "i"
liquid-phase diffusion coefficient for chemical "j"
liquid-phase power constant
(Dfl/Dgjf
gas-phase diffusion coefficient for chemical "i"
gas-phase diffusion coefficient for chemical"]"
liquid-phase power constant
Henry's law constant for chemical "i"
Henry's law constant for chemical "j"
ratio of liquid- and gas-phase mass transfer
coefficients for chemical "j"
f -\
,*• * = v^vf\ -S^- !-
^LAJ IH.)J
Matrix 3: Normalized residual between values In corresponding cells of Matrix land Matrix 2.
Each column and row of this matrix was added to find the total residual to be minimized.
The value of lj/k| used to predict K,_A values In Matrix 2 was used to minimize total residual. Minimum residual value corresponded to "best" kg/h| value.
Units
L/min
L/min
Cm foec
Cm2/see
Cm/sec
cm2/»c
m f(/m gg
mVm30,
Use ke/ki value and
1
1
to solve for liquid-phase mass transfer coefficient (kiA) and gas-phase mass transfer coefficient (kgA)
A-2
-------
CO
•8
to
a
•O
8
c
i
<,
•B
s for Values
1
1
1
W _J -I c c — 1 eJ"
•E -^, •-— e- r~ ~^. Tr
C OJ O) C C Dl —
3 E E ID n E =i
as N
E S 'o '••——•* '
E S n .E -g
JB £2 t 'li= .H
S, » S ra H
| g -E | |
•"S -JS -O >* O 1
s i .g £ i 1
§w S a) -o g
10 3 01 S
9- f- S " S 1 U
-o -a c 9- "c s
— j *^ Q) (ft C ,j
D" O" "o CO JS o
— = {£ en o r-"
.£ .£ fl) .£ **- VO^
O O ^_ Q) O C
f"eo *2 2 To 73
±5 c 5 !f=
fl) QJ co o (D §, M « ^
O O O « O X -o
S
0) o
1 I "S < H " « 1
5od"^od°x> g
3
o
I-
3
C
to
C-~ —-^ J5
"*• - ^
^ cx —
• 3
^ ' ^ -a
oT to
•a
- ^ j
, s (0
t» o g
•° J~* j 9-
1 + i>
1 *<~1 1
•§> w1 ^ 'i
0 ' ^
f~ " ' ^
K <,
1 cj ^
1 "i 1
•C . r^ o
CO ^* CO
•>!?;
s « ^> _.
1 tl 1
» 1 "8 ,s ,
to S J. 1
-^ •= "f- E 75 •*
4 S '5 * "5 i
s> a s .g 1 1
C C C ^ •§ -HE
il«l-5fl 1 .^ .
11 in i il —
Ojfl)OmCOM3C
1 § 1 i i 1 ! 1
sifil 1 If
.= ._ c •— •=; C (o ~^,
C C Q) C CO gj « ^*
"5^&3t"5oS*^x , o
8 1
1 . 'k . "s i . I, . 1
5 o" o™ a cJ ^ a >- x° |
CO
en
O
I
3
W
C
X s g
~ ^
•o ^ S
" ' ' CO
O* 3
-+ o' 1
^ + ^ + £
' •• 1 '^ ^
" — - -> ~ "^ - "ra
•^ W - j CX §
-a -«i- . > * i
1 "X V- ^r>v! '
x: « | J3 o .E
1 • ' ~ i 1
s i ' — r' ^ ^ 5*
^ 0 - * -' , s <, •
! «u A ^ i
i II H II J
-C O
co • « M J= eo
O
2 ! ®
2 . . £ §•
II i !
S || j
£ ° £ "a
< S a. co
c 5;
o ^
1 | r
o £>
in S
A-3
-------
I It
S I I i 11
IJ
a -5
||
£ S
?J- "„«„<, «- S".
u
+
„-
i t
o
O ID
¥ '-^
- 1
<,?•
s
J3
s
•s.
£
Q
i
s]N
^jfE .eflJ -.n
- s s 5ifi
iiiiilH
•PQc:3'tsOCO
Illlllll
I 1 ! 1 1 |1 1
|| | i -» E | |
'
> •?
o
i
n
N
01 >
E 0. S
O
I
O
•n
§
A-4
-------
1 '
w
CD v ^
m '!
c
CO Q «
c ,_ ,
1 t__
1 1 w
w ^H j«
5 [o "*
0) ^="
to ,- N
i
Ii CJ ^
CO Q
•> ^ fS
S ^ -^ ^
ct 1 w
i
CO
•S
> Q rt
CD ' — — »
C v- ->
1 "
0
D) °
C ^~^
.c
CO I* ™
- x "'
of o-
•« 8
CD S
U
8
o
i i
fl> 301
1 ?f
*Q_ "O "O
3; • o g
.< So.
'TS
lit! J. Iti
CO
1
"ro
"o
"iz
>»
Q (D
_. o)
S "S ra
JJ* oo^>r>aojl o
•§
T)
QJ
S.
1
D>
O
^ » X "g
^ >I< i
II M E
II c
X w S
CD
!^
3
< " S>
« > N I
II £
11 Q |
>>
^
£
"^ <"H O -^
^ J HH fc.j4 O
1 - M >- ?
•> + =>.
^ M « CJM y
n O* K* ^ cr
" II H =
N . >. ^ 3
c
0
CD «
J! f
5 CO
*n "5> ^) .£= .£
= E E E E
1
"o
S
1 *
CO „
"ca
'o
1
, - "^
~v
=
o
_x
s*
CD "T 3
I 8 -^ °
i ° I "
cJ o" ^ JL *c
.S" o
§ =P
•a
^
_ 5-
CO (U
1 :* '
1 5 ^
OJ
=
I
3)
A
% (5 T3
co c:
£ *: -
9)
g
^ g ^
1 -s. s
-° o
I a-
1 ffl &
0
i
5
i * if
^ *r « •,
C |
-------
SHOWER DATABASE
A-6
-------
s
tft
€ .1
? £
- &
II
o S
• ™ 03 s z :
fio
- v = S a
2 11 f I
jflll i
p j3
0)
E
ra
ra
o.
0)
M
1
i
Vi
1
1
C
JJ
"n
c
'i
o
0
c
1
s
1
§
(1
!
5
•i
!
t
;
^
<
)
3
3
a
ra
•^
IE"
J
^c"
~
* t
-g
Ji
g _J
c? t
j-i
„ .
cfj
° tr
©PC
cf -2-
0 -3-
CM w
©>*£
a -a
o 1
S)\
X"E
,|
j
O
1 "
U
11
II ^
c
w c
.S.Sfc,
a 8 s s s
1 S 8 S 1
s 2 " 3 S
5 2 S S S
3 * ; « <
" J? S H <
0 0 0" 0' £
o o o o c
ft oi ». 1 £
?> S f
1 1 1 1 1
8 i S iU
T- T- T- T- 1-
S 5 3 2 «
5 3 2 S S
3 « S S S
o o o o ^
o o o o c
| I s s :
3 s S 2 S
3 | 1 § ;
o
X S S g S
^ 3 5 i 1
0
S 1 ^ !9 5
d d ° °
§J.'|1
1 f I 1
^ 3 1 | a
u
c?
a
' 1* * I
. 1" J 1 I
l2 1 1 1
S i 1 2
Jr 1 s i
2
f
0)
2
S.
*ffi
C
1
E
C
|
2
g
c
o
c
c
E
c
c
^
D
7
c
1
C
en
i
•I
!
c
j
<
£
D
f g
1
_.
^
jl
ci
i}
< ^
^ t
-:g
Jl
_
J_J
^
^_,
"fi
Q.
c? |
" 1"
©^E
n-2-
-* tr
N.'*
gl ^
I J.
I
6
I"
U
II c
O '=
s
(0 C1
a. =
(o a
g 5 a is s
in g in w tr
o »- in c«i o
S S ft S3 9
2« in e^ r-
u> to to u
in in *r oj r-
T-- (N <0 <0 U
oj 04 3 fo u
§1111
o o o o ~
o o o o c
II33'
8 s 3 3 .S
°' d § o e
^,-, , ,
u o o o i
^ S 3 5 S
S 1 s? 5 £
CD o ° °
1 I i ^
SI -• a s c
n
2
£
-1 I Jf I
, §• S £ -i
A-7
-------
1
H
2
to
1
c
hemical Co
u
w
t
s
a.
8
1
u
tft
c
o
1
I Operating C
8g
s
s
*!
< i
s
< -^
*^L
-g
IS
c?i
jj
_
«i
II
°
S?
"3
®l
Chemica
fe i fc sg
^ 51 ^ ?: £
CO CM CO CO
T- T- 1— T- f-
rj- CM t- i- CC
CM to "f" ^~ •«•
CO N. CO CO tO
T- CM (O CO
0 °* 0 0 |
o o o o c
CO CO
50 7 o to 2
CM Tf -t 0 «
^5" CM *^ ^
» „ "
bpray lype =
Duration (min) =
Liquid Temperature ( C)
Liquid Flowrate (L/min)
Gas Flowrate (L/min) =
Shower Stall Volume (L
3erson Present =
'
i
J
1
entratio
o
lemlcal Cor
o
a
o
a.
*n
c
1
«
u
o
•=
ra
1
0
2
CO
o ,— .
o s^
ra
s
if
T ?
J
.£.
J§
-g
d" ^
elf
ji
c? j
i|
^ "o*
^: *"
o 1
©Jr
Chemical
CD en co co N
O) O3 O3 CO N
o o o o c
52 o ** t-
^Z O) O> O3
,- CM CO CO «
•*- co co co a
CM CO --el; CM CC
CM CO CO CO Ct
CD CM ** CO (£
t- co S. i^.
'•O CO CO CM
q q O O ]
d d d o' ,
0 0 0 O C
CD r-
i T "? "*. ?
SS82°
CM 10 r^. cc
r: en co r*- c
^ o o o c
0 d d d c
o o 'o "o 'o
X X X X X
•p; 10 T-; Tf O
T- CD CO CO O
CM co co r^
D O CO to C
S £ co
g .S CO 0*
OJ CO CO
CO . CD CO OO CC
CO f^- v- CO T
CM TT CO* JsJ CC
s s s s s
R N ™ CM §
o . o o '
o* ° d o I
0 0 0 0 C
t^- CO
*7 "7 eq iq «
CM •* O CO CO
•^- CM co' 10" T
_ CM U3 N- CC
_ en co i — co
^. 0 0 0 C
0 o o o' c
3 0 'o "o 'o
X X X X X
^ iq -«-; ^ q
v- CD O) CO O
CM f- CO CO _
D O CO U3 C
d d ° ° "
flj i CO
ra § c
3 £ "a £ §
< HI h- UJ O
bpray Type =
Duration (min) =
Liquid Temperature ( C)
Liquid Flowrate (L/min) =
Gas Flowrate (L/min) =
Shower Stall Volume (L)
Person Present =
A-8
-------
2
15
to
0.
«
to
1
o
8
to
.0
2
c
u
c
o
u
1
o
O
0)
.£
I
S
o.
1
0)
o
10
g
1
tj
D)
C
1
c
O
~5
s ~
0 g.
(/)
CO
CO
5
3"
• w 0
g „- Q.
> C m
'€ 2 o- •
CD Q) C c
0. !C — •= •
S ^ i -^
S->, > J
CD" To U1<0-
11 IS
c ra J2 c
te M =i E
1.2 2^
« S ra^
Hill
£ £ S 1 ra-
2 -s 1 £ e-
T> c 0 . £
0 CD S? <2 c
f §31 §
CD S g £ 0)
f* oi 'S ™
»i as
llll!
» g-I|°
1 2c||
c ra c ° =5 ra
CD ^T O T3 „ *-
Q. ** = CD P -R
CD _r CO CO s O
Q. CM o ra « m
-s * -3 o § W
litSlll
g S s s -s -2 s
fl) *3 r* g g) c* 10
Q- .S ra ™ ™ 'E g
•&* ;*s=>|
2 I ra r-o "? E
^li-ls?
1 il^ll
(I) 0) CO 3 3 M «
ra -s i » To M ^ '
l?^s|s
? S!f!!=!
S w ^ (0 <0 (/J 3 S
ra caS(D
1
01
a
o
2
3
__ ^-^
ra
!-
* 2
^1
[c"
<|
—1
1 3
£ 5-
O* "c?
li
0 E.
^_^
cf |
s i
S/ E
Q -H-
0 'S
0 n"
•* s
®"^
X §,
"8
0)
x:
U
fco
=» "
II
II ?
£^ c
>. ~
01 Q
S! 5
"
• •
»•«
S3
O O
;»
S ft
0 0
o d
¥ f
O
T—
X S
t~ x
oi &
CO r*
O o
E
O
2 1
£ 1?
O S
° S S
u
o"
""
2 1 i
ai|
(2 ^ •§
ill
S a s
J 5 d
§•'
II
2
1
=
te
W (
2
t!
E
!
Q.
0
•5
c
1
tt
<§
C
O
1
c
V
u
c
s
*«
jr
U
t
o
g
|
u
1
j=
u
10
c
o
3
•a
c
o
O
D)
C
1
O
a.
O
2
§ -.
0 g
ffl
5
^
"^n
,^^.
J 1
5.
< |
•^ ^3
^
^1
F g
1 3
a1!
_c -J
c? f
^_^
0 ^
d" ?
^
cl-j
o -—
C\l «
@/ C
o>
-------
I
° 1
« s-
5-
«=•£
J
•!l|
•g till
£
S
S&
cfj
a ••
O
£
111! I *
Q. 3 °" S " 2
M o n J a £
•g
CQ
O
I,
•§
'§
Q.
S
I !
i
'C a a. ja
lit J l|f
ilic?ioso51
s
"O
g
s: S.
I-si
> CK in •
1 1
E O
3 ;
s i
a e
Si § 3 i I
-5 •
J| If:
2 2 B i l
5-035,
A-10
-------
s
e
s
n
£
w
£
S
E
c
a
S
mical Col
Hi
O
tfl
c
".£
a
c
CL
"o
£
a
C
£
c
c
cr
c
1
£
5
n
§ g
» " '
•9
&
en
JiC
£ c" 5 £ "o j
i^* c: j5? jT" § ^ '
>, -3 -o TJ EE > !
2 2 "5 '3 « o !
d 3 CT O" (Q ,C i
cv> Q LJ LJ CD co o.
*
(0
1
E
2
£*
U)
c
(0
{/}
w
c
c
1
c
mical Col
OJ
U
u
1
\
tt
"n
£
a
C
c
o
s
c
<
I
(
3
O ^
w
n
s
j£"
3"
<,I
•^ i
*[
< —
^ i.
-g
•a ^j*
o1!
^ "ra
° ^
osi
-- §
° -§
0 -5-
CM CO
@«^
0 §
Q w
"?
CN JO
@"E
d" -2-
0 '«
©3
3? ^,
"cc
.3
S
S
*
•«
*
CM
in
CO
d
o
o
d
^
d
d
°
X
CO
CO
CO
d
0)
0)
"m
o
o
1
c:
O o T- CO CO O O
J= i- CM CO CM 10 Z
C "*"
" II
° ^1L u
» Ifiti
n ^ g^'g 3 .§ ;
>, = -a -a LL S :
E 2 '3 '5 co 6
2
I
ro
Q.
£
10
to
CO
CO
c
o
1
(U
c
mical Co
01
0
s
r
a
£
|
1
E
a
j2
s
c
c
S
t
c
Cfl
I
I
1
5
§ g
VI
in
S
<¥
j»
_
, •— TD ^"O U- 5
2 2 '3 '3 « 6 ff
cu 3 cr CT (a jz '
CO Q D n O co Q.
S2
"§
g
g
a.
w
'«£
U)
c
1
n
w
(/)
c
o
1
mical Co
j=
o
.2
|
i
"n
J
O
(0
c
o
•s
T
C
o
c
ff
c
1
d
£
CD
LO
o 2
w """"
CD
S
j£-
2"
S "i
•^^ ^
*"•"*'
[E*
- - f
< •§
* 3.
-g
•= J-
J t
s -*
1 ?
j-i
E.
" ""
CN «
@N~
s
Q ^~-
Tt CD
CM CO
@«"~
E
a-H
o "1
©3
a? J.
m
1
S
N
*
^
*
CD
CO
CO
CM
§
d
CD
1
d
CM
CD
d
CO
o
d
X
•vj-
CD"
"•"I
i
CD
0)
2
o
1
c
O CD CD O CO O C
^ T- ^r 10 co m 2
3
II „
o"-? ii
Sits,
-^ 2 .2 ~ C '
11 flal;
CO Q '_l J CD CO CL
2
S
£
2
«
CL
l~
i
o
CD
d
'o
X
•^r
co'
o
0)
Q) .
^o
c
| en ^pog o
11 n
°^^-u
ii j? g^'S ^> § ,
sT 2 'i.'i. « 1 !
S Q ~J 'U CD CO £
A-ll
-------
j
i
i
£
t.
1
[
^
(j
i
H
J
C
j
!
1
c
o
7
i
I
O
g
I
C
Q.
|
ffi
O
w
§
s
II Oporatlnq Cond
§
I g
1
S
a 1
ft g,
©5
r~ E
*"~*
1
CO
6
CM
*
*
CM
OJ
in
CO
8
0*
o
CO
CO
d
3)
d
1
o
D
X
O)
d
jjj
5
5
5
si o in cn co o o
2 "*" CM 05 co in H
5
Spray Type =
Duration (min) =
-iquid Temperature ( C) =
Liquid Flowrate (L/min) =
Gas Flowrate (L/min) =
Shower Volume (L) =
Person Present =
tn
<
i
i
a.
1
(
£
tn
u
<
e
£
<
]
c
C
1
G
c
I
-------
2
1
a.
I
%
ra
S
(0
c
o
1
"c
c
o
c
1
O
o
1
§.
s
"5
c
1
Q
O
w
i
<
T
C
(
O
o
1
1
(
°
D
5
o
Sg
A
B
^-
if
< f
^ 3,
-I
*=• g
ii
—
•E ?
o" 1"
_^^
eft
— • "Si
O E
CM w
©"E
0 -Jt-
^ S
CM (A
@"E
o" -ii
o i
N "c
® « -
X ^,
.§
OJ
O
?
+
o
N
b
o
C3
0
CM
0
CO
IO_
C)
o
0
0
T-
X
cri
03
o
£
"g
o
s
c
5
Cj T- CM uj
C
3
11 «
0 ^
2 ""3
ii 3 ^
*C* Q) "ffl
||||
2 S '5 '5
a. =1 .S-.tr
CO Q _J _l
to °
rM O '
£j m Z
*~
II
"c" "
E 3 II
"m* E
1 ? '
CD co i
2
"eu
2
ro
Q-
1
I
<0
o
1
i
3
g
E
o
CO
a
^
|
CL
C
1
0
JC
CO
c
c
A
I
O
u
u
c
\
£
Q
gg
CO
0
0
^Z"
5"
rf •—
i.
, -S s s c §
So J Zi CD CO 1
2
fi*
£
(Q
eg
CL
£
CO
I
E
0
1
C
o
1
0
3
"E
s
CL
1
a
O
CO
c
c
£
O
o
I
1
D
3
n
» '"""'
B
S
_^f
2°
^ •—
j? •§
< 'f
^£. •£
< ^
^ t
-g
Ji
Ot|
— -
^_^
1- ^)
^.
f- ^
-
CM «
/a\ N^
© E
D) O
TJ-
o
CM
0
S
6
c
5 " ^ c? uj R S ;
§
JL"
' I its
^ -| S- 1 •§ •§
>; ^ .1 £ | > i
>. 'S T3-D E g
E - £• '3 '3 <0 5
CX 3 Cy CT CO r~
co Q n 3 CD co a
o
•s
E
TO
a.
1
1
(0
(A
E
CO
c
.2
to
u
c
&
u
E
u
s
0
t
a
CL
1
fl
o
CA
c
o
*
£
(
a
£
<
<
i
u
5
lg
£
^"
13"
^ •—
•^ i.
•S-
-- 'BJ
0 &
^" Q)
CM w
@W~
0, |
CM JO
§) E
©5
X ^,
1
"E
6
«
*
•k
•k
Tf
CO
o
o
i
o
to
N;
.
o
o
°
X
CO
"
o
CD
a>
E S
||
il
^ O h- O IO O O
j| T- CM m m in z
§
11 II
O -£* "
His,
^ -g oS® 1
jjj i § I •§ •§ i
(§• ^ [2 iT 1 ? '
2 ? '5 "5 to §
o. ^ tr cr co £
CO Q I3_Jj CI) CO Q.
2
E
2
TO
CL
1
I
U)
U)
a
E
to
o
1
u
c
o
o
1
E
ffi
U
(A
^a
t
£
ct
C
E
£
U
(i
£
C
tr
c
»
(
i
(D
3
0
5 g
CO
CO
n
S
vf
iS"
<; •—
""i.
< f
< *=
* t
-g
Ji
c cr
oa|
— '
_
3" £
__
-- "a
0 £
O -— *
Ql £
o> o
^* 0)
CM JO
@M
£
o 'I
S "g
x ^E,
1
E
~Z.
=3
o ^-ji^
xllt-il
i f g-l 1 1
,>: ^ ^ ig | > '
t 1 1 1 « §
a. ^ tr IT ra £
to a n ^ O co a.
A-13
-------
o S
c?i
:§
1
i
!
c
i
c
1
tn
i
i
t
<
I
"c
0
c
C
"5
'i
> *S T» TD E S o
II -fflll
|
1
\
C
I
t
tn
i
t
^
a
I
|
•J
8
C
c
t
"ci
c
1
a
C
t_
*E
i
s
CL
C
1
O
JS
0
c
.0
1
o
ra
c
"2
at
a
O
§
O g
S
< 1
-^ J
*!
—
< ^
^ ^
~
«=• b.
JS
^ G
d° ^
•=•
^_^
cJ 1"
s_ 5
o- f
®c7~
E
Q-&
° "cT
i) E
a" &
3 g
©^
—
O
1
o
•K
*
*
in
in
iq
in
E
c
o
Tf
in
^~
CD
§
0
D
X
CQ
CM
O
O
0)
i ra
3 D.
5 CL
5 §
CM^ tp
•t-7 CO
3 T- en o co o c
•; t- CM T- CM LO 2
:: ^
3
Ji.»
O -S- N
11 is'cT3.^-g
ii iiiii
§" ° *~ c
§
a. ^ .y .« ra ^ a;
j
i
I
c
1
1
tfl
(
(
2
«
c
1
•E
a
l
c
I
E
a
t
J
•c
1
1
a
O
(0
c
o
IB
C
o
0
I
§
.£ ^
CO
CO
CO
5
s
< ^
J? J
•£•
S- 1
<1* -c
* a.
pr g
J|
*
£ Q
. °> CO
j-i
•£•
,-j.
c? 1
*I
^ E
d"-s.
" f
a\w
<§) E
Li
35
af .E,
8
i
s
•M
•ft
r^
0 -
o
co
"£
o
CM
CM
CD
°
o
3
X
CO
5
o
c:
i a
i S
D CL
5 S
sl O
%
2 i- CM ° CO
2 t- co in co
5
jL,n
Q "c* "
« III
Ll §i 2
>* "-C3 "u TJ or
? s II »
L 5 CT tr co
CO Q 3 3 CD
8 o
UJ 2
r Volume (L) =
II
CO CL
A-14
-------
Scie
¥
D>-a
.2 55
11
CO
<
"
1
o
0
85
ll
tS
s
tr g.
= ?
eft
0
't?
II
o"
2
0
n,
"8
Mass Trar
c
0
j:
!
1
1
8
r
I
cu
8
i
5
_c
]
C
O1
c
1
(
I
§
,-.
0 g.
tn
in
2
_,£-
-1C
^1
*!
*t
o!i
,^ "&
.§
j|
^ ^
o I1
9 "5"
S 8!
"fM2
@T E
"3
si •%
ei~
2? .§,
1
E
^
O
s
3
S
o
M
O
0
X
O
X
CM
ID
t
C
o
CO
o:
c
ll
^
S
11 la
« £ fe
^ |l
2 £ "3
S S | £
11
1 -E- i
T I. ? t
I ? 1
TI S |
II £ I
(B
Q.
1
i
(A
C
_G
I
1
a
C
1
a
£
O
u
t
\
al
Ti
x
c
t
£
C
c
s
(
t
1
- 1
-I
^
^ i.
tr g
Jf
J|
x~*.
I ^
O £
,
£ ==!
^ -§-
Q°-a.
•^- a)
CM ^
at
o i
||
15
i
6
0
oi
O
•
V
X
CM
in
5
1
S'
3
II
0
2
JLl.4
0 ^ Q 1
11 il
..|i
n
ft i
CD co 1
A-15
-------
I s
111
0*
x "S.
" 8 S °
a- S- E
5 I2
gfii
°-g g
- 0)
Sri
r5 -H.
0
§ •§
0 0?
i 3
-i
lo-^
j-
o5
• <
Sri
S
A-16
-------
2
1
2
(0
o.
tn
c
«
(0
(0
c
.0
?
•£
0)
o
c
3
s
i
u
o
•1
£
S
0.
hemical
O
(A
C
.2
T3
3
ra
.E
2
oS
CO
ZJ
o ~
o ^
a
CO
S
S
jt
^ 3
.—.
S- ^
-*
£ J
P- g
1 5"
. P c
- — '
= ?
o> CT
^^
*
4
*
O _
-
in
CO
o
o
"o
x
o
»
^0
X
in"
m
CO
§
-a
CO
o;
5 o
0 CM CO CM g 0
.g ^ CM CO ° g 2
C T-
!L "
gj fc ._ —
8.1 !{Pl
^ g }^ LL JO 0 ^
t^ 2 "1 "5 CO £ E
:a cr cr to to o
CO Q Zi Zi CD CD a
0
i
S
(Q
Q.
i_
to
c
w
s
(0
c
o
1
•E
o
u
c
3
s
'E
u
u
0
•c
g
S
Q.
hemical
0
(A
.2
'•5
I
m
c
1
S
o
£
3
0 ~
Q 0^
w
s
s
i" .E.
S
'E
CO
s
§
-R
CO
*
CO
CO*
o
8
CM
S
f-
^
^
X
CM
in
3 . .
o
c?
c
| ,= s g' o | i
C T-
" II
0 -£ II «__
„ mi;
Q. 3 CT CT CO <5 flj
co Q _i _i CD m a
2
1
E
ra
a.
1
5
£
.2
1
c
o
o
.1
i
^
o
$
r
0)
a
g
a.
1
E
£
O
, n TJ TJ U_ £ 0
2 2 '5 '5
° £
" -o
CM w
©^
ra o
cfS.
" f
CM JO
@"E
d" -H.
o ^
©3
3? &
S
'E
CO
6
S
•K
*
T—
in
CO
in
o
cri
o
o
s
0
x
f
L
X
•^
in
in
CO
1
Q
&
^
Js 8§°|^
i "-
11 n
£?-£•« 'i
• IPi i
t>-5 !2 12 .IJ- 2 c
HJ D" CT (Q (0 (V
£0 Q D n CD m CL
A-17
-------
J
a
Ji
O -n-
0
i
u
o"
5
" TO i
0
§
5
s
" ^
CM M
II
O"
CD
OJ
O>
s 1
?: x
S 1
t— E
cf-S.
« -o-
<«t 0)
CM tn
«n1
® J
's
"E
JZ
O
•
* .*
< •
•
-R .S ,
1 1
o o-.
s s
0 0
* _ •
O
X S
3 3
3. 3
1 i
ro •§
ca &
g
| -r- CM | S g 1
3 E
II
o'
••— *
1
1" I I- 1 e- 1 i
&B5ZIU
A-18
-------
i
f~
co
" .§.
elf
mr
1 "&
O E
C? j
o -3.
CM W
®*v
E
d" -a
0--3-
S J
© E
-C" O
o 1
LO m
iwi -i,
® n"
"ro
6
O)
o
-K
*
*
CO
CO
S
S
o
o
o
o
o
CO
o
o
d
0
o
rf* •—
•" 3.
-I
•~
^1
=• E
i ?
Ji
elf
.~
o°-|
<*s
° IT
©"t
c;
cT-2,
-<* 0)
CM JO
@) E
c5" i^
o Is
^ "E
?{
"eg
1
CO
•fc
•k
*
S
fe
h-
O
O
CO
CD
Si
CD >
0
O
X
ai
-------
S
1
a
E
'§
d.
£
S
1
1
5
0)
E .
I I
S g
| S
£ c
1 8
J c
C j-
8 j.
ra ,2
A" S
•a
a
£;
S
auld not be de
«
•s
E
2
«
a.
4
c
M
n
c
1
*c
8
c
o
u
I
OJ
o
„
1=
O
n
I
hemical
u
tn
c
.S
I
01
I
CD
§
o
o g
i
*3»
< •§
3
^-,
s!
^
^f
*g
Ji
osi
.&
^-^
J-S
__^
5 iJ,
**•"'
O •;>
•* 8
CM .$
®I
s s
^ E
d" •§•
J 1
||
I
g
J2
O
*
*
•
«
fe
to
d
o
d
-
^
d
I
d
t
^"
0)
d
CD
g
s
JC
"H S §
•a •§, CM m o -g, c
f 1 1
ii "
"sltl
M 3 3--§ •= «
JL. f aj3-§ •£
at ||l|l
If ill ll
n 3 O* CT" W ,r- Q;
03 Q 2 3 CD (0 Q.
"S
£
n
o.
£
u
c
£
n
Q
I
- ffl
C
C
o
U
1
O
(A
01
I
I
1
0)
£
U
to
c
.2
c
o
o
O]
1
o
S ^.
0 g
i
3"
«g- "
•^ ^
, „
S|
*"
<, f
cr g
1 5"
o5!
ji
&
Jt
4 =^
O p
*—
O x~s
CM ,w
®l_
Tj- 0)
CN ^0
§^N£
d" ^i-
o **^
II
s
£
o
*
•K
*
*
•«
0)
d
o
CM
CO
d
00
d
I
^
CO
CO
CO
d
CD
i
1
1
•g c c:
•§ -|, j;j 0 ffl -g, C
S "S "S
we c
11 -II Jl
0 'E'J^=i
Itlliil
>. = -n -o LL. S c
S 2 B B w g «
£0 ^nca^o1
p
I
Q.
O.
a
en
§"
CM
§
¥
1
1
•g
-c T~
ill
..
'g
Study year:
Reference:
Solution Me
.
"
CM"
16
S
1
CO
I
CO
'3
TJ
S
to
g
_ "o
S
O ^
V A
°™cf
Assumptiol
Comments:
CO
1
f
c:
CM
C
CO
CO
*co
o
SP
(D
Q.
C
1
o
1
CO
.2
J
.£5
w
03
1
CD
1
O
CO
1
CO
3
0
to
S
u
=
8~~
CO
"o.
to
0°
CD
a
13
2
=3
Ul
|
II
C
1
I
13
O
3
i
•o
to
£
1
•S
o
2
1
11
•K
j
R
Q_
QJ
U
C
1
i
C
O
1
8
c
£j
S
'c
0)
U
to
1
1
CL
75
.y
1
.1
=5
1
01
c
'•s
O
o
O
w
S
^
-!£
S
-
^*.
1
'5*
J
T
%
^
I
§
^:
O
CN
CO
S
in
CM
CM
0
T*-
O)
CO
X
•^-
T-
o
1
X
q
in
L,
x
CM
cd
CM
d
x
CM
CD
2
,
o
o
o 10
u
o"
**-'
S
"- 1
f—^ p
C 0)
Duration (mi
Liquid Tempi
|o
cn en
CM CM
10 S
co %r
co to
CM •*
CO CO
in in
Ve
x x
h- en
CM -^
o o
"? t
X X
CO CO
^ CO
o to
x g
"*". d
h- CO
o o
d d
V f
X X
in -fl-
ed CD"
"* CM
d d
1
f
1
UJ -C
O -c
0. H-
<0 CO
CO* Tf
•c- CO
11
"c "^"
1 "p
"aT ~^
"S fll
Liquid Flown
Gas Flowrati
TJ- ir
en a
ss
CM T-
h- a
CO CN
co «
r- o
T- *£
- cv
^5
? =
X X
CO O
I"-' f
o c
m
5 E
o c
r- o
en N
^ c
° c
1
f g
O H
|i
It
3
CD
E
2 n
^L. Q
llShower Stall
llParsnn Pres
A-20
-------
2
0)
«
1
£
CO
K-
co
1
» c
>, -B
CO "g
Q. 3
CO Q
co CM in co cc
t*~ t*- tt
CM iq ^=t r*; o
O3 00 O) O) P**
rs_ •— ^"~ *® i-
^ ^. U> CM £
*- ^ o o c
CD
5
£ c
1 £
_? UJ S 1 g
8 CL - js 6 £
CO g. J 5 1
II „ II
II *— ^
O •£' ii si
^-* '^ o* fl)
CO •£ ~s £.
^D 0) ^- 5 1
& 3 2 S !
^ 1 1 S i
f -f II j
s
u
2
£
s
CN IO ^" N-- O
O> 00 CT) G) I*-
^v
^ • ^" "^ C
011 ^ o" o c
0
i
£ E
a) E
E £
•2 2 <
-s- UJ ^ 5 a
O O -c -c C
O Q- £ O t-
CM 3 3 5 C
^ £ s 2 2
" II Ji.
o" ^' n ?i
*— • •?= ^^ CD
0) •£ •=: £
1 a|l'
J" " S
T, -0 U.^ C
'5 '5 co 6
5 5 S S c
2
o
o
a.
i
n
H
0)
(0
ro
s
c
.0
i
u
§
u
1
1
o
r
0)
Q
1
1
.£
o
(0
.s
1
£
TO
c
1
g
s
to
sg
CO
CO
s
°TJ>
•* _
^i
c
^ 3-
n
^ s
sr g
? 3-
" *5)
0 ,§
„
o5^
^.
^
3 'ti
*-*
s~-
.5~" D3
-£•
• \ ^__^
^ «
@<
o CM to in c
CM' co" in h- cc
o o o o c
t ttt a
•*" •»- T- T- u
X X X X C
co o m CM c
co •* iri cri
^ o oo «j ^
v v -O- O S
x x 0 0 c
T^; co Q Q c
CO v
r«- r*- oo 2 f*-
0 O-O T C
0 o 0 <=> C
s o o o o
t— f— -e— T™ v
X X X X X
CN u? T h- en
o oo ro cj N
_. _ . N
X cq g $ £
N "*" o" o c
CD
CD
2 fi
J- LU ^ § g
0 Ix 1^ 0 K
CO N; 3 OT C
^ o> co ? 2
II „ jl^
o" -^ n d-
"*~* "£ "^ ^
CD •£ •= E
| o"3-f -
^ 3 si s
T3 -a iT S
'^ '3 « 5" E
cr CT co x;
3 D CD 0) D.
c»
g
1
CO
w
. "C
5
1-
*s
f~
1
I
ffl
£
•B
8
CD
i
,3
1
C c
i **•
•5 5
1 I
1 i.
I So
o ?i£
« «
a> u
* >g
Q •§" ffl
to OT a:
"
en
|
CO
•g
CO
1
1
1
I
•s
8
§
i
Q
I
i -*•
"a j=
1 I'
1 |^
III!
STUDY:
Study year:
Reference:
Solution Method:
CO
*J-
a
&
N."
o
to
CM
>
X
1
h-.
TO
i
'§
"j
I
1
8)
^.
q
~j
33
LJ
CO
" "ra
i-
Assumptions:
Comments:
1
|
1
§
|;
2"
o
"§
CO
1
g
CD
CO
CO
.a
J
•o
I
1
E
tS
II
1
to
i
ID
0
Iwithavc
c
1
0
•o
* = Could not be
I
£
1
£
CO
CO
m
E
§
1
O
'C
o
o
1
E
s.
o
1
U
Q.
1
1
,p
o
|
1
c
c
1
CO
O g*
to
to
CO
**
•*!
DO
O
CD
E
CD
2
1
i *
1 ?: $»g
i
u „
og-j^
^ & c=
Spray Type =
Duration (min) =
Jquid Temperature
Liquid Flowrate (LA
Qas Flowrate (Urn
ll
ii
=i
S
Ehower Stall Volurr
IPprffnn Present =
A-21
-------
e
i
§e>
t
Q.
1
C
[
H"
w
V
I
,5
c
«
U
•jj
t
E
e
C
as
i
o
£
_
jj
U
W
a
•o
c
o
U
c
c
s
E
S
O
2
s?
f g
ro
S
.-
2"
< S
3
^^^
^
^1
«=• g
Jl
°5i
,«.
§ ^
trf
^
o|
S f
CM jS
@w^
c
cf —•
° t?
Si u
CT-H.
J S
J "E
®
o
CO
§
o
o
r-
X
O>
*-
0
1
E
o
2 T- CM u, CM 0 C
2 T- ^- "" in CM Z
3
" II ji.
0 •£•» =i
**** *5 *c* ®
jl. s^^-5 t
iitJiil
t- g H-u. o w c
- I l-l- S ^ ^
7 Q jjj (D co ci
t
(
c
E
2
£
j
c
2
i=
n
u
c
n
c
c
I
c
c
o
0
"n
c
1
a
O
w
G)
1
a
o
£
S~"
1
.C
0
U)
c
o
IE
"5
1 c
o
O
O)
c
1
U
a
O
S
1 f
0 g.
(U
_
jii
«^!s
•* 3
"c
^- 'E
^-x
^1
«=• g
JI
c?!
*— • .
!? ^
cJ" ^
0=1
a?
^ £
cf-^-
O -5*
CM ja
©"e
£3" -2-
O 1
t ™p
®>
c ^,
"to
1
6
«
*
*
to
CO
CM
•g M
II?
o .£
O
CO
•<-'
^
8
o
d
0
3
X
CO
s
§
(U
c:
6 to
d.
P o
O Q-
?i-
^"co
3 t- CM u, CM 0 C
2 •<- "3- •" -, s TJ T> is: s o
S 2 = 5 w g K
o
£ «
me
a E .2
° c-
§ I
3 i, «£ A « C
I— ^ O O (0 O
w w as w < o
f
•o
o
U
s
4
!
£
1
{
(0
C
(
(0
c
c
1
2
c
C
"5
c
E
a
C
«
t
s.
s
EL
"re
c
I
CJ
s:
U
c
0
1
Cl
£
2
a
o
(D
Zl
S „
O ^
to
S
S
^ ^
^ t
-I
-I
-g
!i
o £
elf
^J
1 "a
£
—.
£ =d
°"I
S f
@e?~
E
0 ,,.
•g§ 0
i" ™™$%*
3
jl.ll ji.
O •£• n si
— 'c "^ Q>
H 3|Sli
i _c a) "Hi
a
O
§
|g
(0
s
S
«L -^
^ t
s!
"^
^1
p g
ig
cJ" £
o1^
*£•
_ ^--
1 "&
° -§.
c ^
°""I
si
@c^*
=, 1
0 —
•cf 0)
CM «
©li
Q" &
' — *
^ li
©5
;£™E
"H
.9
6
*
*
•k
*
00
03
S
E_
"£
0
r-
ro
cJ
o
o
CO
o
o
(O
X
03- .
in
CD
O
(D
C
3)
=
1
5
—
o o i*. 10 o g e
2 CM CO CD ^ CO 2
5
£?•£• ii 3
n HII "
1 ^ m "?n m a
S~ 2 "5~% wo §
3. ^ cr cr ro ^ S;
co Q n j CD co a
A-22
-------
- en .to
>a)g
I'
'i.
2
'E
ID
E m =
i • s u s a «
i d" O < iS C3
S S
§1,2.2
h— jj m O
CO CO
.
O to
CO <
I
o
•a
s
I
0
H
j:
c
i
•§
0}
2
'5
§
£
1
•8
.e
i
£
i not be de
H
C
U
II
*
s
o
•s
2
to
a.
0
1
1
1^
w
ffl
s
c
c
"«
1
s
= JJ
I
a
.£
(
t:
(
c
"
<
I
(
c
.
:
t
c
£
1
O |S
I
s
^"
b)
*f ^
b _c
. z^~
s
j- E
"•^_
* .—
*Ji
^ £^
^^
1 "3"
J £
c 3-
o5|
J-a
~"
£.5
0 •£
O -—
®t
O-o
CM jg
©"l
o "]
©5
of B,
Chemical
c
1 °
c.
II
o 1
^ c
i— o
>» +3
2 2
Q. 3
CO Q
O CD "d" (O *^J" CO ^" CO ^1" 00 IO CO CO
CM-^^-^^-^.-tO-
o'fela'^lsfbtftsfbt
xxxxxxxxxxxxx
ococMSocMor-^i — oococ1
OOOOOOOOOOOOO
fllllllllllll
T-T-CMCNCMCMCMCMCMCMCOCO^
OOOOOOOOOOOOO
o"oo"o"o*o"o"o'-o'o"o"o*o
OOOOOOOOOOOOO
o'ooo'o'o'oo'oooo'o
OOOOOOOOOOOOO
xxxxxxxxxxxxx
O>O3O5C3CJ)O)O>O)O)O}O5O5O
CO CO CO CO CO_ CO CO CO CO CO_ CO CO C*
o'oo'oo'ooooo" o o c
I§^<2l2^*2*2's<2*2l2 2
oo o o o o Q-O.gj5.g-p-
0600066666606
CD
o I*-; ri c
-»- "^
=i
is"
•* 1
^~
* •—
,- J
t5~
o, 0>
o .§.
£ 5"
0° f
, — .
Jl
"""'
cf |
CM ^
Si
Tf Q)
©"E
C3--H-
o "\
|i
Chemical
c
1 5
3
II
a. -S
£* c
ra* |
Q. 3
CO Q
CM 0> 0 CD T- CO
CM T- CM CM CO CO
ft> fb ft
X X X X X X
CO CM CO O CO T
T^ ^ v^ CM* CM CO
O O O O O O
Illlll
CM CO IO 0) O) CD
CM CM CM CM CM CO
O O O O O C
o o o o o a
o o o o o. o
O _ 0. 0 O O- 0
fkttt'fc
X X X X X X
OI O) ,OJ O) OI C
CO CO CO CO CO C"
o o" o' o" o" o
o o o o o
sss sss
5E3°||
JL !( —
O "c* " ^«
3 — J t 3 II
•gj _J O •*-*
' S. 2 f"l.f
I™ ul o <^ ^"
T3 TD Ul > O
3 13 (0 O" *2-
CT CT (0 .C 03
•jj ri o co a.
A-23
-------
oooooooo
xx
g g
xx
"-g g
-
co co co co to
II
o
"
• <• i '
\£ jo
=5 o
Q) U
c .
°- "8
3 ™
"a ®
§ 0
£ 1
a) (a
in pj
c .S
E =>
Q) ^
CD "to
| 8
§ g
CO ^, (D
01 ^ x
«- § «
s •£ °
ai • co •—
1 » 1
I 11
111 "a> i_r
° M °-
z if
1 1 '§
O g 8
a or =5 " E
1 g- i £ 5
ro S .c — c
1 I Jll
•5 £• jg o s
1 -8 -I 3- f
°- 1 I ^ "'
Isllil 1
° 05 CO ^ (/J 0 £
£
"S
«
I
1
•o
1 Si
.. Is .. 1
S a S S j2 ^
j; §,= 5 ft S 3
Q >, e | § E £
w co or to < o «
;
•
<
(
C
'tf
a
£
g
a.
*n
C
'S
.2
±!
O
O
.1
2
0)
a
O
i
o g
w
I
<°f
"i
<• ]
v^1 J
^
p. £
^3-
S ^
O ,£
= a-
o" ^
•s-
Jj
5 i
•~-
*J
01 o
a —
" 8"
1 jn
' n°
@|
S
(D
§
0
(
ppray iype =
Duration (min) .=
-
co ti- co co *
CM CM CO tf "
.-^-~
/98S*
+ •* co CM m
CO CM CO CT
N CO C>] CM C\
§T- CM IO If)
O3 CO CO CO
... _
•^- T^ - co co
o o o c
o* do d
D'O'O'O'O
X X X X X
O CD tq p -^
T— OJ CO O) O!
[ CO xj- CM £?
03
C
c
p "<
: i a. H- £
?cosi|
1 II
J •=• II II
Liquid Flowrate (L/mi
Gas Flowrate (L/min)
Bathroom Volume (L)
Person Present =
:
•
c.
^
(0
<
c
1
c
Q
Chemical
»
§
s
1
o
D]
I
O
1
O S;
ID
(O
i
<, 1
•£•
C
S1 -I
—
tr g
tj"
t
O ^.
^~
ra 1
_E_
J|
Osi
^
S'J
-* E
cf "-H.
0 -S1
rf
I.J1
®J
1
I
a
o
(
ppray lype =
buration (min) =
o h- co
CN CM CM ^- O
T- T-
-------
1
E •§
€
o
z
E
s=
I
iil
•a i!
&&
'5.'3.
. a. Q.
'£ '5
£ «
ffl C ..
; s o a
i—
to
o
o
§
o
to
to
CO
<;
B I
•§ s
"c§ S
S |
"S "
= ^
ra jf
•8 -S
f~ (jj
o J2
i 'g
1 1
•I
•o
c
S
0
•o
1
1
•D
3
O
O
II
*
to
Q.
tfl
C
4
0)
i
g
_g
1
c
c
o
u
"n
u
1
a
O
|
a
c
Ou
Tf
c
I
a
JC
O
c
t<
E
u,
c
1
(
1
u
J?
"S. 1
"* 3.
it
"" 3
5
c
o
•g
UL
c c
>- CD
5 .> p P P P c
s »iiil5
c o
11 II
o 3- ii «
~" "g "E" — 1
! 5 ! "aT „
ii .c g^ 2 S 'o
g- ' — ' g _o ^ E [
I — Q 1— U_ O O
to 2 '5 "5 to •^•-
Q. 13 D" CT OJ (0
co a j ri CD m Q.
"S
CO
i
Gesell and Pr
STUDY:
E .S-
> to
If
s reported in paper. No d;
fficient estimated using re
CO Q
Ii
|si
fl-
5
1
to calculate K..A or mass
-a
fi
JD
8
N/A
Stripping effic
Liquid-phase
Not enough d
Solution Method:
Assumptions:
Comments:
ot measured.
E
II
1
I
vith available d
^
•o
fl)
c
i
5
* = Could not be del
2
1
S.
£
i
li:
&
CO
0)
c
G
2
"c
1
o
Chemical Properties
t
^
c
<
O
u
c
3
(
5
0)
n
n
^"
<
^°
<.
"*
M
p-
1
d"
I
o
I
o
p"
o
•*
(N
©
0°
HC@25C D,@24C
unknown
Spray Type =
g
£
^
^^
"E
5.
[G"
3.
g
i
E,
t""
»™.
"3>
£
1
"o*
CD
_W
"E
—
'o"
QJ
u
4
co"~
S~
._
o
*
CO
*
*
*
g
o
*
*0
X
10"
c
o
1
C
t1
u
o"
Duration (min) =
Liquid Temperature (
II
II
•H-"
Liquid Flowfate, (L/mi
Gas Flowrate (L/min)
Ii
u
Bathroom Volume (L
Person Present =
A-25
-------
:?a§S:
1
Ii
!^L =
la's-5
I OT K (O •
f
CD
C
(U
D)
S
H
1
I
.to
l-s
1 I
o '01
Its
3 B .2
flf S
i i? 1
£3 fe
<§•!> <
•sis
^ S H
i| s
o S -^
5 8 €
111
Stn'pplng efliciem
Liquid-phase diffi
Not enough data
Assumptions:
Comments:
•o
S
3
in
n
Q>
"o
c
II
"=
;h available data
T3
1
«
* = Could not be del
!
i
L
i
i
]
«
\t
i
i
(
1
£
0
C
c
c
c.
~
t
S
Cl
.E
o
(0
a
r
a>
o
c:
a.
|
a
0
n
§
5
•o
c
o
0
D
C
a>
a.
O
£
o — ^
O a>
to
3*
-------
BATHTUB DATABASE
A-27
-------
CQ
8
I
s
CO "^
i '
o s
t *
cE S
c *i:
a «
» *CT
I t
3
.C CO
^ 0 £
•S ™ £
a | T>
l_ O (D
Si? -1
S 1 §
|i 1
I* 1
| | 1
.g s S
2 •§ §
S S e <
S *^. S v^
*S i* a) *
-S ia O T3
o: o co e •
|££ £• 3
'g T- CM -R "°
S TJ -n ° .2
m « s M a ~™
SSiiii 1
TO
_p*
1
1
1
.. o
« -a
*3 O
». "S c .. o
MI fii
*J £ H to E ^
A o o e» o u
co a: co < o *
j
<
i
TO
c
t
•i
c
(
tl
E
_£
1
1
C
C
1
a
§
u
a
r
0)
c
CL
|
J
u
to
=3
c
o
u
ra
£
o
a
O
|g
I
f
"^ S
*I
=i
<, i
sc 3
^"•^
«=• £
JI
«?*
elf
o5 |
o -—
••* s
CM .to
D. §
a *—
•^ "o*
©-E
iS -2.
o "B
x°1=
.1
6
en en co oo <
CO CO CO CO t
SCO h- h- if
CO T- T- •
V" ^- T— 1^ T
O) ITS CM xr; T
CM -^- CO CO C«
T£ S °1 °^ T
0- Q' CM CM 0
tp. ^ co h- co
CO CD CM CM O
OS CO O) CM J
0000^
d d d d c
o o o o c
S-rt- co en f
CM co" CM" ^
CM in en en j
in CM -**• co c
CM in r-. a
31 en oo r- tx
*l 0 0 0 C
0 O 0 0 C
*o''o''o<'o'9o
X X X X X
^ in -r; ^ q
T- en o) co en
§T~ O CM CM ^
. °- d d «
o o
S * o
•§ s «
2 " co o o
5 _ m -0 "c
f S 5 E c
i
— co CM **"• in *^ ^
j? CM 0) c-i £ Z
2
11
3
a>
a1 ^ ii S '
£ -^ "c -3
3 ci -^ £ 11
i, 2 <° 3- m "£
*C* O "ffl 0) 2a 1
i 1. t i E s s
5 IS 1 1 w 1 i
a. 3 tr cr ca Jr te
o Q n 3 CD ro a.
(
1
I
(
c
tj
(
U)
(J
(
10
c
c
1
t
a
t
c
c
s
E
c
C
a
|
a.
8
i
o
o
"&
1
O
O3
c
V5
IS
Q
o
o g.
ca
S
"" 3-
co o
CO CM ^- CO T-
o in •* en T}
^t- CM c
° d o" o d
=> *0 ^3*0 *0
X X X X X
^ in v; «^ q
•*- cri cri co a
x- CD
~ ^3" CO O ~*
O 0 CM CO °!
°. P d d tt
o o
-2 c g
0 § y >
5 *^* c ,^ "r
5 jr -1 1*1
< DJ [2 uj c
5)
3
= a'i CO *~ S *^* C
F CM 03 CO t-2
L
II
s
o*-^ n >
(P t .5 ^
3 •£ "^ 1!
' E c" 5 2 -2 ^
|l£ ll^ *
1 1 & i » i i
O Q 13 ^3 CD (O Q.
I
i
!
1
c
J
1
c
1
(
i
p<
1
c
c
*ci
1
c
m
c
a
c
c
a.
i
E
a
s
U)
c
I
o
o
en
1
g
o
fg
(O
ro
5
.g'
J?
<^ S"
^ 1
s-
.5" -5
s
<_ 1
^ 3
<=• £
Ii
o ^.
0°|
^-*
1 ?
cJ i1
x
os j
° "cT
®(C^
£
o" "^
CM i
@)^E
a &
o 1
©•>
xotk
1
6
D
D
D
L
11
c
I E
II
5 2
1 3
0 Q
en g 3- ^ CM
en °. cfi co co
CM CM CM CM CM
"* "* "* "* ^
OT U) CO CO x-
CM CM T- ? i
o cn co N. x-
eo' -* -^ •* iri
55 CM, iq in x-
o- -r- -^ ••* m
w. T- co en co
LO *- CO CO CO
to cn r- CM S
co in co CM K
o o o o ?
d d d d c
o o o o o
fs. T- v- CO CO
CO CM ^ co ^
cn ^r f*- cn co
co CM to in oi
CM m h* a
~ en oo i*-- oc
. o o o c
0 o o o" ci
tf of ty tp of
00000
X X X X X
T— IO V- ^- C
^— OJ O3 OO O
Tj- O
CM CO OO CO
o o co in o
°. °- d d T"
o o
•§ * c
«, ^ c >
« y o s a
< ffi |2 ffi o
8-8g4
T~
ii
o
=
Ji. " •§
0 S !L>
§ 3. 1 J „
s « S ™ -E
|| || |
£ £ c | g
^ = to o 2
3 3 CS to Q.
i
c
J
1
J
i!
to
1
(
S
(0
c
.i
1
•g
u
c
(
c
7
'i
J
W
a
'•£
a
c
2
a.
"n
c
E
£
o
(0
:!
T3
C
o
0
cn
1
a
O
3
CO
I8
f
, — .
D) E
3.
< -^
si
< f
3
(T g
J!
•- ^
O g
*""*'
^T
O"~ £
cl-j
a |
t^-i-
" ?
SI to
5" ••— *
_5 "^S
®«^
DC E
1
1
o
1
i °°
5
E
ti
I 'g
I |
o a
iisss
CD CM CM C>
v- ^ CM CM C
o o q g |
0 0 0 0' o
0 0 0 0 C
o -si- o T- 55
•, § §, 1
< UJ [2 jjj c
a 5 1 1 5
ii
^
II "5
O •£• II >
'i'l | i •
1 | 0 I I
J" w ' " i |
CO Q.
-D S "- § 0
f fill
A-28
-------
(0
o>
£
2
CO
0.
'M
I
to
td
J
1
c
u
o
U
15
1
(0
.£
i
2
Q.
8
I
CO
£
O
to
c:
5
1
ra
1
C
§
O
(0.
to
ro
^
j£"
^="
!
<"-
^ s
<-£
li
° £
,^
-f5 CT
-S-
_ ~
O" E
' — '
-- 01
^ ^
-1.
O ^*
CM ^o
Q ••—•
O "^
" "E
©^
X ^
8_
E
CD
6
.c
cn
1
£ oo
IE
li "
» I
o c:
S3 O
CD 2
O Q
T- m
o o
o o
m uj
T- in
T- CO
CM T-
CM r-
•^ CM'
CO CO
d d
CO o
CO OJ
8S
0 0
o o
CO v
CO CM
O CO
Tj- CM
^ CM
"^ O
0 d
0
0 0
X X
T-^ CO
in -*
CM OO
So
d o"
CD
S
o
s >,
Q) £
|3"
-i s n
_J OJ +-
-rg §
ff S t
2 -2 £
s
U- & 0
ra 0 S
CO ^ CO
IS .
CO
s
'•§
5
to
CO
X
H
o
• £• .-- - .;
'2
' 1
c
:D
c:
.a
•§ •
cj
g
5
d
^f
CL
.v;
•<
v.
•§
S
1
Oj
,c .
JC
c
S
1
"s
'£
O
" 13
i
cn
-i 1 1
o | „ |
Illll |
5
3
^D
g
' §
S
0
•S o
o ft &
"5 c .. o
gliltli
§ •» & 2 g.l «
1— 3 o "o to o u
CO 0)
CM CO
O 5 CN CO "T
§ g ° ° "
CD CD Q
"" "m o ^
f l-g| J
"5J fr ^ fr ^
^ LU H- H C
u:
TJ-*-»- co £2 ^f
CM OJ 1^- CO £
II
5
d
^ M ii j
1 1 § | - 1
<0 JO > CO =
§" S S 2 i
15 S — ^- -
'S" n" E o o
a
2
(Q
a.
1
2
•-
CO
E
rations
i
§
u
1
E
O
CO
a
P
emical Pi
f
u
»
C
^g
^
I
CT
.E
1
o
1
_ . — .
to
to
i
S
2" ^
—
•^
•«-
CM CO o CM h-
m in co co ^t
T- in CD ^t- I?
co ^ co CM t;
o o o o g
d d d d c
o o o o o
__ ._ If
o co p oj a
CO -^f CM CO C
•^- CM CD to
g J3 co S a
8*J ?* -^ ^" °c
& ~o y* >
-------
o
a
1
a
O
J
. §
CM ,ra
S«*fe
acf-S-
10 in 10 in
co a> co CM a
o 10 at o £
co CM r- CM m
o> co, cq o Tt
10 co -«
in. h- co
i- CO CO
0 T^ ^ CM (0
CM CO CM "> ">.
o co «. -t
• t- in
cn h- CM CM ^j
c
CM CO CO (O Tf
-,
.e£
«
W
0)
(—
"6
1
c
o
*•§
•tr
0)
(0
Q
Q
1
1
fe 1
g>1
•i &
'£ "a
,§ c'S :
111
O Q) CO
| ll
•Jg •§ 8
Q£ *J TO
.§ -2 -2
| i | u
gisi
M
TJ
J :.-
Reference:
Solution Met
Assumption!
1
•o
0)
1
s.
based on j
?
CM
C
'
i 1
s i
^
s
TJ
I
U
1
5
Comments:
* = Could not
<
i
j
t
0.
k.
•S
(
F
(i
b
(
i
1
q
i
o
c
_t
1
o
'•E
a
1
CO
U
to
o
S
£
U
CJI
2
o
1
o g
CO CT
T- CO
T— tj- (O __ _
0 0 CM «•? 0?
q q o' o co
o o ,
(U ® m
8 1 g S |
^ ffi i2 i2 C
to
S£g|!
n
S
(D
1 —
^—^ "o
O II >
" „ £~^
•g 3lf ~
till!
A-30
-------
js
S
sfer Paran
W
1
o
1
O
u
c
o
o
"ra
O
|
c
1
s
E
2.
o
£
O
S
1
E
0
O
m
s
}
S
0 g
^
•-. =d
° £
^^
J, i
*£•
cfj
" IT
CM .$
o
Q - —
O -5-
•& CO
CM .w
Q" -S-
0 li
CO « CT
CM Q
o">~
X £
i
6
i"
ii
i— * c
ro "5
0. 3
CO Q
en
*
•K
0
^
N.
in
CO
o
o
O)
CM
'CD
X
p
*
=9
i
sl
< -i
^ a
^g
li
ii
£
^~.
J- 1>
—
*i
al
S) w
CB O
Q •=•
O -jf
^. §
CM
00
*
_
0
X
10
CO
*
§
In
cc
^o||
11 II
"si"
uid Temperature (
uid Flowrate (L/mi
s Flowrate (L/min;
throom Volume (L
rson Present =
cr cr 03 co c
Zi Zi CD oa a
*
Q , = 73 -0 U.
1 1 11 a f i
CO Q Ll "J3 C!) CO C
<=•£
0
s
A-31
-------
KITCHEN SINK DATABASE
A-32
-------
H ^ CO O W
eo w Qi co <
,g
t?
1
'o
$
§
8
I
•§
1
f=
0)
omment
o
.
•o
S
1
determined
J
1
2
1
II
2
"S
2
CO
a.
^
n
1=
I
to
c
.0
§
c
o
o
75
o
£
.O
CO
'•E
a>
a.
g
a.
1
Qj
O
to
c
_o
^
c
o
o
O)
c
•c
2
OJ
Q.
0
§
1 ~
0 g.
CO
s
3*
ll_^
5»l
£
^ i.
•£•
^ i.
-g
t"
_
oSt
f— »
g =d
0"" ^
^_,
el j
° •$•
CM S
@CM"
E
0 -3-
•^- CD
©*?
E
d~~
"•1
©j%
^ -§.
o
'ir
^:
O
* * *
III
co co in
CO 0 CV
CO p CN
d *"' *~
CD ,- ,)
Tf CM C\
o o c
o o c
in S °°
"* d d
8 £ £
ZI CO CC
^ o c
0 CD d
LO CD -J CC
d
a
c
a
X
CD 0 a
S <5 c
III
CD 01
C CO «? 0 c
z CM ,f -^ 2
II
•£-
II <»
11 II £
o E-F
^^ 'c CD
CD -5 °
•i -1
o g^ 2 S
1 I ll J
fe a- cr >» .«
CO
0 g
CO
CO
CO
S
S
^___
$i
"*— '
*i
«."!
* t
<=• £
i ?
0°" .§
c -3-
d" ^
iff^
1 ^
^ c
__^
£ =!
0 |
"•s-
•^ CD
CM CO
@w~"
£
Q°&
° IT
M" Q)
CM .«
®» .2
< "_j LJ X C
2
2
m
D_
>w
to
C
!
E
to
c
o
I
S
c
0
U
75
u
I
£
U
a
• "€
.CD
a
S
a.
I
£
O
to
^
c
o
o
a)
£
1
g
0
cu
u
o *~.
0 g.
CO
to
S
S
x— ^
S|
SI
< —
^S
FT g
i 5
S CJ
~
cl-l1
^ _ ^
I "^
^— ' C
^^
^ ^J
o" g
"-5-
"^ CD
CM CO
®c7^
E
Q°^-
o -5-
c\i S
®<^
E
Q -2.
o 1
S "E°
©^
af £.
1
E
S
* * *
CO CO C*^
5S-S
O N. 03
T-1 d d
CO "^ °
d d d
I^ CO ffl
o o o
000
CO S c?
o o
O "^J". ^
1^- O) f-
m cc
5= § g
0 d d
? *9 "9
° ° £
XXX
t- t- o
-*-1 co a
O CM °
d
0
c
-* "E CD
CD ^ O
II ^ £ ^
O CU CO , .«
n "_i x c
2
1
n
1
U
to
E
!
S
CO
c
o
•^s
1
§
n
0
o
"(5
u
1
0
'•E
a)
a.
s
a.
1
E
o
to
i
'a
c
o
o
O)
c
1
s
0
§
to
gg
CD
CO
s
s
-, C
^» ^
-^ :§
~*^
-!
< -^
^3.
pr g
•g 3"
u. Q)
o ^.
c J
6s 1
^_^
1 =^
^ c
^
= ^
0 1
° •3-
CM w
@w^
E
cf-2-
0 w
Si |
@N_
t
C3T -H-
0 1
CO ra
© *^
X 5.
1
E
6
* * *
CM CM 0,
•^ N- 00
"^ CO CO
» .5
<_n 3 x c
HI
A-33
-------
2
]
£
(1
CL
£
10
c
1
to
w
n
c
E
|
c
o
C
|
O
S3
I
a
8
a
1
I
o
c
.2
"D
§
U
I
1
a
a
O
2
o .— .
O g.
S
n>
s
f
^,
^1
^
S- i
<, |[
16 a
cr g
Ji
—,
tit
"si
**••«
cf|
o «— -
C7^-
0 "1
3? .§
L
i
o
3
i
*C co
S> CM
•^
m
rType =
Temperature ( C) =
1 1
•* • *
CO CO CO
V- ^ T-
*~ ^J- CO
CO CM C\
cq co CD
§ ^ ">.
d
* CM
CO •«
in co
53 co o
0 d d
•fct-t
XXX
-1- -r; p
T~ O> CJ
|§3
CD
C
a
||i
< R O
>
II
*£*•
•g
rlowrate (L/min) =
lie Residence Time (i
S 2 «
3 x c
A-34
-------
D)
-a
o
* "fo "° -g* ~ •§
llllfll
.. *•
S SJ S .2 J3
3 •=£! = « E
I— ^ w o in o
^^
elf
Ji
^
cj- j
ZL "cT
•^ "^
°"E"
"^•^
'E
o
S-
c:
.
II
o"
¥
8L 1
^ 1
S -o
2 '5
Tj- co 03
* * *
* * •*
* * *
Tf in f~
co in °
00°
-^-
o o c
o o a
°og
CO CO <*
CM CN CN
o o o
in in in
XXX
CO CO CO
W ^ T-
CM CM CN
000
000
O 0 0
"o "o "c
c c c
ro to c
UJ UJ 5
S ^ 8
~ |(
H
•9 c i
__ .5 '
'5 . ^ J
2
tJ
g
ta
(Q
a.
CD
U)
ffl
£
tfl
(A
ra
S
(A
|
Is
u
c
o
o
1
g
o
o
.2
a
a
g
a.
"n
•g
£
0
c
c
*
c
U
Ul
c
1
0
£
3
CO
og
CO
ra
S
==£
-^"
'n'
ai "E
Jtf 2,
"c*
^- "E
^
si
p g
1 S"
S |>
^-^
elf
if
"'I
O ^«
CM CO
©N£
Dl ^
^r co oc
S- CO h-
CM CM CS
co co a
h- N- N
CM CM CN
CM CM CN
o" o a
? *9 •?
O O O
XXX
CO CO CO
•^ ^ *•
* + *
"o "o "c
C C £
ra ca n
UJ UJ ii
-------
WASHING MACHINE DATABASE
A-36
-------
o
h- f
3
§ g
CO
S
^
>
<• •£
* ^
<- 1"
V ^
""*
5!
tr g
d° ^,
t "ci
£.
li
* — *
•— j-
t*~ — '
_
"o*
Xf Q)
CM JO
@) g
a"-^-
° ¥
N W
©"E
ca -a.
Ijr
X ^^
"S
c
"£
^
g> g g CO
T— V-
* ££ ^
. ^- O x- '
CM CM CM
o> co cn
CM" CM* CM
R CO N. CO
2 T-" T- CM
«- oq in 10
CM cn a> •*"
I ill
o o o o
to cq p fm
>n CD CD CM
to in r-* in
w [C CD" CM"
in N- co
II co r-- co
T 0 O O
° CD O CD
•9 "9 *9 °?
O O O 0
X X X X
t— T— "<3- O
T— cn co o>
cn
o d o ®
0
c ffl
m Cg
N S
ffl a, g o '
O § £ O
1 III
Q)
CO
f co o, 3 u, 5 o
t^
1!
II || o>
"?Ji. i
§li!"
11 15 ~cu 5- y
it 'c' - o3 73 (b Q.
o 1 & g 1 -g £
0) 3= T3 T3 11.
"o S 'a "3 , 3 CC CT (0 C S.
O Q 2 IJ CD LU. C.
2
1
s
(0
0.
w
J<
1
£
(/)
(0
i
(A
C
G
1
03
O
§
c
8
I
a
^
I
1
8
E
u
j
(A
C
<
A
1
1
O
o
jy
I
1
(
i
£
5
u 8
cn
CO
2
^
2°
c
a '£
^ ^
< 'I
.M -£
^^ s
^1
pg
If
_
— =^
£,
f^
1
* — *
""^-
£ -J*
0 i.
* 'cT
^ E
Q" ^
" -o
>i jo
cafl
o 1
® Jf
"
"cc
c
'E
(L
§ § 00
. «?«.«?
., in rf in
CM CM CM
•* co in
in in in
oo CD cr>
CM" CM" to"
^_ co co in
§ S 3 g
O O O 0
IO CO CO CD
m rC CD t-"
to in co t-;
W CO I^ CM
m h- co
7~ co I*-- co
^ O O O
° d d d
"? T T *9
O O O O
X X X X
V- T— -^ O
T— 05 co cn
§ S S §
o
1 «
g g i I -
§ S S o
a> 3 g* o
""I
< I
^
^!
tr g
= ?
s' •"&
* '
_C 1
° £.
t"
' —
-- 'a
•^ a>
N 42
c?-H
Q-I
0 1
3) m01
0™g
C
E
(U
•5
2 CD I^- CO
^ O3 OD [*>•
, ^ ^ ^
CD (O (O
* -<3- in in
( 0 -r- 0>
CM p in
*^- in i^
I"; CO
cn
§CN C^ °
82®-
d
g o
g ffl i |
1 S €.S
^ J2 ffi cj
1 '
* ". cn ™ in g i
_g co *- ,_ « ,- c
S
n
II n a
o -H- ^ §
"m* E .£ o
11 ti'I s-
[] C CD (0 Q) O.
a> '£ p S 2 -o
CL "— * £ O > CO I
,>• "g |2 E I « „
t-= -a -a u. -1-
SS '= "3 M ra :
3 cr cr aj c ^J
o a "13 n CD u, c
0
E
2
£
w
£
g
assTra
S
10
c
1°
•-I
§
0
' 8
1
o
"1
g
0.
8
- E
£
o
i
i
g
c
CT
c
•JS
i
<
i
O
0
D S"
8
3
^
^
^^
•£•
•* ^
^-
*I
-£•
.^j c
3.
^
p ^
•g ^~r
° -§.
= 5
cj° f
1 ?
O" E
£ i
0 ^
si
-1.
^J- 0)
CM _W
cat
o 1
JM "I
§) "tfr1
X3 E
J
E
.E
0
&is|
* * * *
* * * *
CD
co m cn CM
o ^j y,; ^ *
o
p CM o °i
CO CO ^~ CD
CO _,. CD CO
CO 2- h- CM
o o o o"
o o o o
^ CO CM O)
m r^ co* v-
^
O '
|| .C
-------
&
|g
«*
J
Jl
in ro •*!•
in TI- in
S =>
N. CM CM
S S3
O O O O
O O 0 O
M OS CO t*-
U> CO CO -r^
0 0 0
T" T- V
X X X X
^H> " ?Z
5-g- N §
I
o S
i
5-1
o o o
to" m" to'
S S.
co m co 10
CM co co -^r
CO CO £2
o" o" o §
o o o o
CM -n 0 CM
in •«• x-=
O CO -r-
co co oi
£; co r- co
^ "o o p
odd
X X X X
T-; T-; -^ o
*~ cn co en
io en
CJ O *
m t- m- °
5 co 5 50 jo
35SZZ
1
C
1
>(
I
1
1
1
)
^
M
C
Jt
1
\
c
o
C
~
1
fl
£
O
to
.S
1
"•
1
X
U
v>
c
JO
^=
1
Operating Cond
£
5>
G sS,
en
crt
ra
S
^
5-1
6
CO O •*£ CM
co en co h-
^ iq iq io
CO CO CO
*• S 5 8
co ^ m
to eo" co
CM CM TJ- ^J-
0* -* •* co
2S Si CO
CO CO CM
a II CO CM
^.00
d ° o o"
o o o o
o <° ^ en
10 **• « o
v- ^f CM CO
u^ co to T-
10 t^ eo
ZI co r^- co
^000
0 0 0 0
D O O O
X X X X
T— T— -"^ O
T~ en co en
o
T- Tf CO _«.
8 S 3-5
o
1 1
5 1 -S •§
el m §
i
- K •*- ^ in S o
5 5 CN cd m ° c
^
^_,
Cycle Type =
Duration (min) =
Liquid Temperature ( C) =
Liquid Flowrate (L/min) =
Gas Flowrate (L/min) =
Final Headspace Volume (L
Clothes Present =
i
i
<<
i
!
s
t
C
(
1
Ti
<
c
c
(
t
7
<
1
c
^
Q
C
CL
1
a
U
c
o
~
Operating Cond
£
D ^
w ~
CO
s
3-J
~
^ 1
< T
^ a.
sr g
1 S
, *" D
O £
li
.^
^-^
cJ I1
o —
cvi §
S) g
•4s1 *Si
O —
^f m
M w
Q "^
n rt °"
0 °\-
— •
Chemica
CO CO t- CO
en m co co
^ co o en
* O Tf- -d-
^ co io in
CM LO h- IO
T^ CO CO ^J-*
°. ^ "^ CM
co en en T~
d °" d d
0000
r*. cq co •<*
en ^f en en
•«* r*- io i-
^_ IO 1^- CO
^ o o o
0 d o" o"
(0
D O O O
X^ X X X
t-; •*-; ** O
^~ co co en
§ 5 3 <°
o
. 2 0
S 5
lilt
< £ Q o
1
- }2 en m. u, o o
_g 5 T-
-------
>
§
a
q
jr
a.
° U.S.
(o c 5
111
o •«:
"P -tj T3 ~a
Hill:
•D
O
.. I
n ai S
-^g-Ss
If
Ss
5
ra
•a
^
J3
0 1
c w
§. 5
•i
TO
g
1
•a
0)
w -*-
go
A =
'ag 'I
'E 1 g
S 1 o
^5 ."
g
"a
£
iS
(0
Q.
1
n
en
55
CO
c
o
23
DH centra'
c
c
1
o
—
t
Q.
f
t
£
(
ji
o
- i
.
;
t
_
I
i g
i
1
^
in
l^
' ^^
f- E
5-
"
; ^
— 0
5 g CO CO CM
•* ^ ** ^ •*
*"i ."*^ *"": ""^i .
o o o o o
H5=a».s
o co co r~- 1s-
CO CM _ ^ -^ CN
•<}• CO
P P t^ to CM
o o
in ^™ ifj en en
i- co « 5
§§55- §
CO CO §i $ CO
P P 0 0 P
0 0 d 0 °
r»w in u> c^ N
CO CM CM CM P
o
T^- f~- T~ "^ *°
it CM [^ (^ CO
CM in r— co
ZI oj co t**- co
^. O 0 O O
0 d d o o
if ' t?
3 3 ^3 S i •
(D ^ CD O. £1)
a. E 73 _w a.
t£ ^ ^ S c '
1 1 « g I
2
c
TO
E
n
a
1
P
u)
B
1
1
1
C
O
1
0
5
t
(
'€
c
1
t
i
s
•
:
i
.
i
~
1
i
^*
n
-!
*c
<• "E
*i
_^_^
S1 '^
s
^g
Ji
^i
O^ E
c?i
0 -3-
i 5
(N .w
O°-ii
o -5-
TT 03
©^
iS-H-
O 1
en « ra
«5
3? £
1
c
6
I e
^~
i
u s
1 j
S 1
ff ;=
^ •* CM ^ CM
° ^ ?I en r--
-p- -r- v- T- T-
T— T— V— t— T—
r- 10 - IT
•^ fl)
N w
Q 1-
o 1
CO n "
^ £
x° £,
"c
j
O
5 *"
It
» |
S. -^
.>. c
5" B
in co v T-
GJ en en co
J* Jj JJ ^
d d d d
O T- CM Tj-
«- t*^ cri co
CO «5 CM T-
x- 0) •<- •<-
g in CN ^
D. -t- CM en
o
P CO CD CM
|>^. CO CO CO
VJ
2 o o o
0- d d d ,.
3 C3 O CD
o" o o o
.. _ CD
en "^; P co
CM CO CO 0'
T- CD en o
CO ifj -^ CM
_(_ m r^- co
° d d d
? .
-------
0 g
«*
D
O •
1-
Q.
T
1
he
§O1 N- O
h- K. CQ
ra co co ra
o o o o
383
l/j CO CO t^
^
5 5 g
n m o-
in r. <^ m
" U5 -^ c*>
X X X X
v* oJ co c>
? •= «
•i 12 S
n
0
*MlSli|
J 3 O E^O
|J
(f&
°?
.
©
co tn N-
f*. i^. TJ-
CO CO CO CO
d d d d
CO CO CM CO
SSSKg
d d d d
iss
o o '
X X X X
•^ *: -t °.
^~ OJ CO OJ
3 te
°
2fl
S «
•g §
Q
i
1
I
1
Q
il Concent
P
^t
j
o
V
e
£
a.
"n
Chemic
i
1
cS
%
*3
S
o
1
Q *~.
W ~
ca
S
S
if
SI
*i
<=•£
•Ji
"5I
ell
o£j
° •!?
N 40
D) EJ
3
D,@24C
(cmz/sec)
o 1
3? £
Chemical
,_
° 0> CO I^
CO CO CO CO
CO CO CO CO
•^- °~ co 10
CM j: T- CM
OJ CO CO O
CM T-" f^ CO
IS =-3
Si?ftS
CM CO ^ ^
O O 0 O
0 O O 0
° d d d
CO K. CO S
CO CO CM Q-
T- CD T- in
^ CD co CM"
- o fe o
0 o d d
X X X X
T- O) CO. OJ
d
« P $2 «
CO CO CO CO
CM CM CM CM
CM " !£ ?2
r*; W Tj; O3
r*- CD m" CD
OJ
d
T co m "3-
OJ IO CO CO
d d d §
£ CO CO S
°. § o o
0 d d d
ss-5s
CM OJ CO CM
CO m" ip' CM
3= g fe |
0 d d d
X X X X
"*-. •*-. "f. o
T— OJ CO O)
||58
5 ^ j=? "o
2 o o r-- S CM § S
n
U :f
2 ~ J
f.. S -T -, ., II |
1
1 1 s 1 II I i
§1 jfl <§!!§
1
]
^
1
5
j
1
1
c
o
c
1
I
c
. .c
o
u
a
t
c
a.
Chemic;
(A
T3
§
o
D
e
a
O
§
o ^-*
0 g
S
5
S
s-f
sf
s|
•=• £
Ji
j|
Ji
o5|
o — .
N CO
©"E
cT-S.
0 -VT
•^ 01
"^
a" -S.
o T
1!
Chemical
° § 0? S
CO. CO CO CO
O CD h- >^-
cri co" id CD"
CO CM CO O
W CM CO* **'
is g ».
| d d "
S9!SS
T- 05
§si§
d ° ° d
d d d d
™ •* CM 2
in 0 oq x-_
^f 1^. m CM
^Sfeg
° d d d
X X X X
v OJ CO OJ
1
1 81 SIS
0 0 0 W
o
UP
0 ^ COOJCOO^(/
g T- -' tO -"
11
.3
G" ji^ i
"*""" .. c "o
O II '5 3>
ii | 3.J. 8 -a |
3.1 lilt J-l
f | £|-|x| S
§1 ft Jill
A-40
-------
3 „
£
j£
O E
a &
o -s
CO CO CO CO
co -i- P en
- - T- co •*-
iri CO CM CO
i S3 S
S 5
0 0 0
. - .
° 0 0 O
r~ in. cq
cc i in T- '
X X X X
5 CO §
IS "i
ei"
g
^ -S ^ 5 E E 1 J
f § Us s
CJ Q =j =i CD ii-
*
»
5
o s
g g CO W T-
° ^ 03 CO CO
. . .
CO CO CO CO CO
CJ CO 00 CO
CO U> W CM CO
CD CM CD O
U3 CM •*— T—
g g g x
o o" o" d «°.
0> CO C^i P 10
^" ^~ CO CO t~
«- en co £: co
o o o o
X X X X X
x^ in T-; -sj; o
5 S S «. "i
0 0 0- 0 CD
fit
UJ F UJ
-
CD C C =J 2 TJ C
fie I Il
CO 2 CN CO CO
OJ ° CO CO N-
«=•
CM CN CM CM CM
O ^ 03 O> CM
CM 0) •«- •«- ^~
00
CO CM CO t- "I-
TT m CM co -3-
o o o o
o o o o
o CM T- CM •*
O O O O
0 0 p O
d d d d
X X X X X
*-; tn •*-; -^ p
•c— C> O) CO Ol
..
5 o m. -t
p o j-; 0 to
o o
u '• •= fe g a . .
i s II 1-8 5-
g^PSil §
=
l
!o-JL
IT) IO W IO
=>. "! -
CO CO -
«> jn co en in
d a « CM •*
" SS 8:
Sqggg
0 0 O O
X X X X X
*— eri cri fld oi
S S CD
llil
B ?• a ?•.
S S 2 S
I" 1*
t*= T3 TJ U. -S
E = = (0 "TO !
ZJ CT CT CO C
o Q 3 =i o C
og
< *
i
^ cn r- t*: oj P
CO CM
-------
1
E
!
c
Jl
I
J:
1
f** — co r* CM
cn p to in cn
S S S S S
o o o o o
S S S g S
S S S5 §
55 §5 g §
o « 2 2 °
s s § § x
<= <=
s *
0 0 0 <°.
O) CM fc H
Si i- <=>. o.
o o
.
0 0 0 0
X X X X X
i-t to *•; -d; C>
v O) O> CO O>
3°
uj |
! 2
e
I
I-
,@24C
m2/sec)
X ^,
CO OO CO OO CO
CM' CM CM CM" CM
O) CM CO CD T^
CO* CO* CO -^- CO
in in 01
- - *»
T- oj r*- ^ co
5 0 0 g 0
co -tr 10 co co
o o o o o
CD C) 0 O O
o o o" o o*
. rj- C*- CO CM
co CM ^ co" T-:
T— O OJ
i
0 ° "
„
2ocou>cau>(/i
= T- -r- CD U3 to ,(0
•§. 1
o o
-
&18 I
3 3 O iT
U)
i
j
c
\
1
1
1
(
£
(fl
c
_c
|
c
C
Tt
c
'i
o
__
^
c
s
0.
1
I
ffi
o
(A
C
.0
1
o
c
S
§
i«
o §
CO
1
1
jtf
•£•
S
s-t
^^
•^
^1
pg
J!
«?i
j-i
<*!
o —
CM jj
d"t
^ CD
S ^
^ E
o 'g
g"E°
<§)•>
0 "
x --
1
1
o
=, °
u
1 £
f|
CM co LO T- 0
CD CD O O CD
CD CO CO CM CO
O) CO ^ CO CD*
O CO CO CM CO
cn co" -a-* co CD"
CD
co ^2 10 N- cn
go - - «
IO CM CO ^ CM
T- CM CO CO CD
i 1 § § |
°- ° o o o
55.11^
0 ° 0 CD X
IO
3> CO CO CO ^
LO co *=f T- r*-
•* CM tD > If} T-
^ cn co h- co
*^ o o o o
° o o" o o
o L "fe L t
X X X X X
V- (f) ,_ ^- Q
T- O> OJ CO O)
° 5? <° r- CD
^ ° CD* "*" ^~
CD °
^ £2 •<»
lllll
^E^mS1
gS|.Ii
It .
3
II 0
5 n |
~v II -S 1
a o~ ^ o " jl
| "oT~ g^"S 5
— 2 ^- El '-5 j||
o
•g
<3 £
•5; Q.
So
| o -5
^ 'S S
I » 1
s o * . £ 'S
| |||| J
w ? ^ s c? ^ £
I
•8
.. I i ..
1 S •» f i
Ifljll
co to o: co < o
)
1
1
c
ij
1
(
<
J
E
_€
£
1
c
c
"n
E
S
u
„
s
I
i
1
I
U
c
o
S
c
8
C
1
o
Q.
O
1
0 g
TO
S
5-1
s-l
—
*!
jr g
JI
"•si
-=<
s 2*
O P
—
«*!
*«*
s »
CM ra
.—.V [^~-
© £
•* Q)
CM JO
* 1 Its
O Q ^3 D CD
IO f~
LO °
II "
^, ^t
I!
o (i
0) M
U) "c
S tj
ID 1-
X 0
— en
C ff
in ^
A-42
-------
S 3! G> C C a _O
S. ™ ^. E. S —
\ «- B. § -a 1. *jj
sfel^"
1 1 »._j
| ii S
1?
ro 01 sS -^ to -». §
o en Q -t». -*• o v>
O
§
o
to
•t*.
CO
X
o
p
o
0
o
3!
o
0
'8
bi
*
»
O
)
5"
3_
"3" ^
^ (§)
3« S
Q
•cT P
S~~~ K)
4^
•2- o
~JD
c? N
•2- o
if
|f
p i-
"^"^
!f
3 ^
-g
1 ">
fim^
^ P^.
i. 1
3-
?"«r
5" -1*
y?
"2.
^
:
^ 1
' (i
1 ' -
!
•>
|
i
)
3
J
i
>'
n
9
9
5'
a
O
o
3
D
|
5'
?
M
3
U)
•n
TJ
01
3
3
CD
CD
55
: H §. 1 1 1 ®
g-pgl ^
o 5" •-
I """ -9
n
S g S " o w
o
o"
1
p
K
CO
X
o
4
p
O
'o
o
8
o
o
CO
0
CO
Cf>
*
*
o
3*
>'
J
=Lt?"
L (§
3W w
^ o
• — * c
i©
§^~~ N)
•t^-
-^ o
\»
rl
-2- o
ci P
=» 5"
If
£ 5'
If
^ - o.
3 •*
E^ —
^
5" ^*
,^
C pj-
3 **
^
i--^
3"
77
°:
e
i
0
CO
3 g
(
0.
1
1
;
o
rr
i
3.
i"
i
3
5
)
3-.
>
Jp 3 ^
' »•§. f •§ 1 ij
8 "5 f* §• "
1 <: i. Y s
3 " 2
"P
II
» a i -s s s |
1
&
3
p
o>
CO
X
-*
A
p
o
o
0
bo
: OO
0
o
N
to
*
*
O
J
[
~3
C. _
| >
"F5
— • >
P-
**:
1
8 i
i
-ci
i
;
'
•
o
3*
)
3
I
p
)
j
0
p
1
r>
3
0
3
1
n
3
u
in
ii1
1
TJ
a
3
3
1 ^ III | |
ttf CJ CD 3 ^
CO 3 §"-0 ^ *
;"g o 1 S 3 "
S 8c-p| « .
i f i " 5.
3 " £2
II
oagSBs'l
o
o1
3"
3
p
01
CO
X
Q
*
p
o
^
o
o
.'g
o
0
B
ro
*
.
0
0
5'
a
"3" iF
U ©
3^ fo
L <">
•oP
3,,®
CD 4=
•2- o
--. O
3^,®
I? £
•2- o
3" 0
P= =
f^^
|s°
3 c
B ^
if
3 J
i ^
S
^
1 >
§•=""
5" ^*
55-
°~
|
9 Q
— ' C
c
c
1
'•
•)
r!
0
r
>
3
r
i
•o
3
o
}
5*
n
O
D
i.
v
s
3
5
3 -
i
3
01
0)
CO
31
1
•o
Q>
S
3
CD
5
CO
s 3 o cr. c o o
3 0) ja .n c: *<
to co c c g s.
~ Tl Q. Q. g: ra
•|i||ll
o 3"
I ""' ^9
ii
1
o*
p
CO
X
p
o
o
0
CO
ro
o
o
d;
*
- •
*
o
D
i
f-I
fl
cn" M
(D •£••
•2- 0
o-P
3 /s
w* to
££
r2 3
|f
CQ cp
pS 5-
«^ °"
i ^
if
i-5
S-
•^
E
c
3 |
c
»
:
)
•
O
3"
)
}'
)_
3
I
)
)'
0
[
i)
3
1
9
|
n
01
w
w
_l
ransfer Paran
(D
ro
a
-------
S 23 G) C C O O
3J ZJ 01 JQ .Q C *•
3 ifl^S
s? sr *"** *"""*
a 3 " 5
5- "
a c
iT "
s • • *s » f
m m n
* * :
CO CO CO
XXX
pop
N> to IO
3 ^
s
n §
c_^
If
J5
f
s
g5 Q
1
f
e:
Q
c^
Q.
5*
3
Chemical Properties
o
=r
>
^
3^
•)
>
3
)
i
I
r
i
3
Q
•
1
r
F ? &S"5* ? -9
J J T1 51 S g: <
O 0) S O <& ^ "c
5" S* S c ^ "3* <
o || f 1 "
n g*« — — •
i I " Q
~ ffl V
is
IT "
«
-^ * * * 0? <° g
hi rn r
03 CO 5
2. 2. £
CO CO CO
XXX
CO CO CO
XXX
^ ^ c
pop
ro ro N
CO CO CO
Vj b> ^
CO CO CO
1* en Vj
O) _», ^.
ro o o
* * *
» * *
* » *
» * *
SBS
Chemical
L"!
CO" M
(D ^.
-2- O
*o"«P
So
5|
->. M"
2 —
tu
w
31
g"
=• S
2'-=»
^:
- ">
^
|>
I">
^^
S
3 —
i
o
n
CD
u
5*
n
O
o
Q.
5"
CO
O
5'
•o
o
0
at
3.
r
o
3-
O
3
5'
•>
o
3
)
to
[
f.
;
I
!
r
* O > cn 3J cn cn
O 3 2 c 3" o.'C
||| |J|S
a" 3 i s °
S> g. 0 | 3
o § S
3 ^- CQ
3 S §
8 ^ a
<» > a
a ' i*
1 I
0) O
3 Q)"
o cs
8. -§
Q. <•
S. a:
I -^
O) CO
IT CO
§o
•o
QJ _A
Q. ^
m it
8 g
a
CD
D)
3
2.
(D*
5"
=•
tn
D)
;*
CD
1
(D
3
o"
3 .
Cycle Type =
Duration (min) =
Liquid Temperature ( C) =
Liquid Volume (L) =
Gas Flowrate (L/min) =
:inal Headspace Volume (
Mass of Clothes Present 11
3 .E - '
n «
(*• cn « -g co -x n
ui tn g -^ o o a
0
c
c
£
o
K
110
CO
X
c
p
c
b
o
s
o
0
=5
^
*
*
*
i
I
3u"
\ t
to O
L. o
•2- o
•Q-P
3io©
2r M
2 *•
•2- o
If
JI^ •
If
^-^,
ts°
^* c
£a
3 *
='
I*
1?
^
DJ
CO
CO
S o
I
i
Operating Con
Q.
5'
m
O
T
i.
a
1 Properties
o
>
>'
^
o
>
i
Q
[
£
1
}
1
P
,
Cycle Type = "
Duration (min) =
-iquid Temperature ( C) =
Liquid Volume (L) =
Gas Flowrate (L/min) =
rinal Headspace Volume 1
Mass of Clothes Present 1
as
iT »
^ cn « -g lu -A n
01 cn g *. CD o S
O
c
c
S
"*•
CO
X
c
p
o
0
bo
to
o
0
00
o
'-g
*.
H-
*
o
zr
CD
i
3«":c
% ®
to w CJ
l"f
•2 o
fr D
fs,®
CO Iv)
CD ^
-2- o
I-r°
C 3
If
,_^
if
*""
? o
it
•? ^3
"5 7\
Is
m^
S=
S
^ o
— cT
c
S
Operating Con
|.
5'
3
O
3-
D
3
1)
1 Properties
o
>
i"
L
-
>
3
>
»
1
R
1
i
i
i
-------
Sfr-V
-------
5552
- P P
CO CO CO -^
o •*.-*-*.
X X X X
- E P
o> OB A.
o o o o
Q (fii to CO
25 -*4 o> -NI
en co w tn
-* w o _*.
CO CO (Q (O
-4 W ® (0
-6
li°
Io
IS
,f
if
jP
g-3
•£
ptf
»•
t- ra II o>
|£ o
ii « 7
-X ° °
CO CO CO
"o '**. '-* "-*
X X X X
co po ro
CO •%] CD
o o o o
« CO (O **•
g 00 -N! 0>
S S3 2 S
LS
ITPH
.-P
If
if
^?
§
3
0
5T 2 Q> -Q Ja cr «<
1 1 f| Ej f ?
3 -^ £j> E" 3 "g" "§
1 § ^ 1 1 .5 "
f*ri si ii
IT 3
aiS o
n ii n
•j
C? 5 2 2 M 0 |
CD
l^ § §
O 2 CD (p
® 1
ro ""J •£* C
01 +*• ex
ro
CD 00 CO -^
O i* ^* tj.
X X X X
o£ ql C
en ai At 6.
P P P O
O O O *
CO -4 CO ^1
-*J -i. _k -O
IA M J*. Ji.
f gg£
O O O O
P 0 0 0
Is sals
|S8£
-^ CO CO -F*
co ro o KJ
<£•>***
* * * »
ro a> ->. jx
1
1
I,»x
S "
s^ o
g /S\
MkQJ
W* M
-2- o
Ii
— ' O
If
"^ "
if
If
!f
2 J
If
^
if
s
i
,.
1?
f
en
0)
•£- g"
3
O
™
§.
Q
O
o
i.
1
tn
Chemical
"0
3
1
O
r
3
_
3
S
D
s"
3
0)
J
>
i,
3
|
1
i
|fg ff = §
III II II
S -1 S c* 3 "31 "S
| ^ <: | ro ^5; "
"• CD" ^- -^ Si ii
3- ro I! ro
s£ "o
n ii IT
_
0 Vl 2 i>. W 0 g
Q. «< c" 2
| | m °
S i
- p p §
CO CO -fi- C
o co a
CO 03 CO -A
O ik. 1* ^J.
X X X X
o o o o
go a r.
"*l CO ^
co -vi tn
en -i. -i a>
tn -* 01 to
g P -° *
-
o o o o
§ p p p
Q co to en
't^ CO CO ***)
en -*• co o
O3 * * *
* * * *
CO 00 CO CO
en 4*. o en
O
f
1r®
L o
f -°
N,®
Si" ro
CD J^
S, o
S" w
•2- o
1-5°
' — '
If
"-*
If
**" "
if
i*
if
_^^
I5
5 ^
3- ^
"^ .,.
's
D)
„ ™
~ 8
o
D
^
5'
Q
O
o
1
o*
Chemical
•o
1
A
8
r
3
5*
n
>
1
3
\
*
>
t
t
3
|
I
['
I
5T
o*
iT
a
m « S.
w 5' 3
Q. g1
8 =
| -g
"a
M -6
S- S
3CT
_ m
l'§
-------
a < x c en o o
to' 2 a» •& -o c: *<
itf**i*
3 S. 8 § 3 •§• o
ff ill? "
^ CD II CD
3 .-5 -~-
SCO
v!s2s -
O < I C C O O
55' CD CD .a ^2 c «<
S- ~ K a s. 3 a
£§:§ » .
'2. |S |
^ CD II CD
~ -^ T
lilts if
ilt
5 *
II S
=£
S v »
O < X r~ •" D O
S' ^ CD ^' ^' c •<
lilllis
S 3 ••
c
•=•
1
^ .
H 3
S
S? <" m "•" * -'if
8 -a S i» -
-2- o
(O 00 CD
b !p». '-x
X X ' X
pop
bob
co -j co
co -^i cn
-2- o
-c?-0
3
tD CO CD -x
04^^^.
X X X X
3 *.
CD W
•2- O
CD "*J CO 1O
158*
o o o o
_ _.
"ro tn co
O O O O
P o o o
f
!f
o o o o
o o
KJ w
.4*. &0
o o o o
P o o o
if
If
1" ?•
si
• P P w
a> <° -^
p p p p
§ § s s
K3 4*. CJ1 4^
f
If
CO CD (O (O
10 to -s] -&.
3g
O CO CD -C*.
SCO
CO
*
en -u co co
-
*
9 Q
SCO CD
-4 M
I- >
1
*
Q
- 8
sss
§ CD CD CD
o cn co
-------
81/-V
9ICKW-101-OS9-100Z :33lddO ONIlNIUd J.N3WNH3AOD "S^Hi
a <;
entilation Rate (L/mi
ishes Present =
n
if £3
o
0
5°
1"
i
3
CD
b
X
°
o
b
CO
00
00
en
o
b
en
o
o
b
CO
ro
o
en
en
en
,.
*
00
£>.
J.
eadspace Volume (:
n
co
rn
jjjf
cr
ro
a
3
CO
CO
^
X
—L
o
b
^
^
p.
^
0
b
8
o
p
CO
%
S3
»
t
*
00
c c: o o
ycle Type =
uration (min) =
iquid Temperature (
iquid Volume (LJ =
Q.
u
<
2 '8 .1 |
— Q
c 2.
CD £>"
•
—X —X
X X
3,3,
P 0
0 -.
oo ii
en
en o
P eo
TO m
o o
8 8
CO 00
o o
fo i.
en oo
CO Ji
w en
01 jo
* *
* *
* *
00 CO
00 ^
O
ro
3
I
-
g I
•e»°
|.@
D3»a
(Q w*
L. o
* — * c
3 g)
1 c
— C
"cT to
(B1
Co" ro
CD Ifc..
o
-2- 0
I.P
^-
1--°
£ 1
*^"^
o-
gf
"3 o
si
3 J
1 ">
3 ^
S
^~.
i" «?r
-E ""*
y..
"^.
~
0>
CO
CO
""^ n
-2^ 5"
CO
c
3
Operating Con
Q.
E
o
IA
O
=r
f>'
E.
TJ
1
§_
«
in
O
i
5'
O
3
3
I
j-
n
a
n
in
^j
3
3
(A
31
TO
3
3 '
i
3
D <
entilation Rate (L/mi
ishes Present =
^
ii
CO 5
Cyclohex
0)
D
ro
%
CD
CO
b
X
_k
4,
O
b
00
00
01
o
b
o
b
g
0
b
00
o
o
o
s
0>
Jt
X-
g
_L
eadspace Volume (;
C
ii
CO
§
!
§
ro
*
^
CO
iw
X
_x
A,
O
b
^
0
o
ro
en
o
8
CO
o
fo
CO
"**
s
*
^
#
CO
o
c: c: o o
ycle Type =
uration (min) =
iquid Temperature (
iquid Volume (L) =
£)
u
^
2 S °> |
Acetone
Toluene
0
P b
O) O
ro en
CO -i
LA LA
X X
o o
i, A,
O Q
b "
oo ii
en
CO ^
CO -P-
-pk
0 0
b b
3 8
o o
ro ix
CD CO
co eo
•^1 .~4
CO ,-tx
-*• eo
* *
* *
* *
8°
Chemical
__^
3 i
M°
•^ (§>
3
ZJ O"
(0 " Ol
{""
~ o
3 1st
_ioAB
ro IS
•2- O
X— O
O (Q
M TO
c/T ro
CD IS
jo
" '
If
'*-*'
*3* ,-»
S1-
"3 ,0
1 Q.
"3 J
0**
f-*
1 >
1*
3
i->
* — '
y\"
"P?
S
1
B §
CD
Operating Con
Q.
5'
3
(0
O
sr
3
o'
SL
TJ
i
n
=;
(A
O
[D
3
o
S.
o
o
3
n
n
3
|
5'
3
CO
U
A
CO
3
A
31
TO
1)
U
3
iT
3
-------
-------
960/00-U/009/Vd3
008$
JOJ
sssujsng
•jaujoo puBij-i}9| jsddn
8t|l Ul SS9JPPB 8ljl Ol UJP19J PUS 'J9AOO Sjl|l AdOO JO 'l)OB)9p
: Q3U3H >IO3HO syodgj 959419A|aoej o; MSIM jou op noA ||
•J9ujoo puemja;
jgddn 9111 u| ssgjppe 9m 01 ujnjgj pue 'Adoo jo i)oBi9p
'|9QB| MO|9q 914; UO S36UEL|0 AjSSSBOaU ||B 9>|BUJ eSB8|d
09170S 'O'Q '
v |E;u9
JOJ J81U90
-------
|